Traffic study for project Conexión Antioquia - Bolívar

Transcription

1 Traffic study for project Conexión Antioquia - Bolívar The elasticities found have been kept constant for the complete analysis period, because the values found in the regression models were acceptable and below the ranges found in other cases. The following table presents the elasticities used for the analysis period for the three types of vehicles: Table 6-4: Elasticities of traffic with respect to the GDP used in the traffic growth model Lightweight vehicles Buses and small trucks Large trucks Source: Steer Davies Gleave These elasticities define the assumed relationship between forecast GDP growth and forecast traffic growth. GDP progression The percentage growth in Colombian GDP listed in the table below used for this project was supplied to Steer Davies Gleave by the concessionaire in January The information sources are the followings: the projection series published in January 2017 by Latin American Consensus Forecasts the long-term projection series published in January 2017 by Latin American Consensus long term trends Projections in line with the long-term inflation rates provided by the Central Bank of Colombia (Banco de la República). Though Steer Davies Gleave is not an expert economic firm, these trends were compared against Colombian official sources, and we consider them to be coherent and reasonable representation of the Colombian economy behavior at the date of this report. Table 6-5: Colombian GDP growth projections used in the demand growth model (March 2017) Year Annual GDP growth % 2017* 2.4% 2018* 3.1% % % % % % % % % 2027 to % *Forecasts for GDP in 2017 and 2018 were reduced by the national government after the date when scenarios were modelled, this may be reflected in lower traffic growth rates in selected toll plazas as noted above. Source: Information sources described in previous bullet points Beyond 2040, it is assumed that the annual growth rate for all vehicle categories is 2.0%. This assumption has been made due to the uncertainty of long term projections. Steer Davies Gleave Page 101 of 259

2 Traffic study for project Conexión Antioquia - Bolívar Demand expansion The average hourly demand between 06:00 and 18:00 was used in the assignment model for both weekday and weekend trips. The results are then expanded to calculate the AADT for each segment of the project. A demand expansion model was developed based on the analysis of historical traffic data on different time spans (day, week, month and year). From this model, different expansion factors (EF) were generated for each toll plaza, vehicle type and traffic direction, linking hourly flows with daily, weekly, monthly and annual volumes. The following figure illustrates the process to convert hourly traffic into average annual daily traffic (AADT). Figure 6.5: Traffic expansion process from hourly volume to AADT Source: Steer Davies Gleave The AADT for each toll plaza is obtained by converting the network model hourly traffic sequentially into daily, weekly, monthly and annual traffic. The annual value is divided by 365 days to calculate the AADT for each direction at each toll plaza. This ensures that the AADT reflects annual traffic seasonality. The figure below shows a worked example from 284 vehicles per hour to an AADT of 4,518. Figure 6.6: Traffic expansion process from hourly volume to AADT Source: Steer Davies Gleave The expansion factors developed account for the seasonality in the traffic during the year relative to the values obtained during data collection. The following are the expansion factors calculated: Average weekday hour to daily weekday Average weekend hour to daily weekend Weekly to monthly (during the data collection month) Monthly to annual. Different expansion factors were obtained for each toll station and are summarized next: Steer Davies Gleave Page 102 of 259

3 Traffic study for project Conexión Antioquia - Bolívar Field work: information collected during a full week of counting during November 2015 at the master stations associated with the project s toll plazas was used to obtain the hour to week expansion factors for modelled traffic at each. Some toll plazas were not installed at that time (Manguitos, Caimanera and San Carlos), so the nearest master stations were used as the following table indicates: Table 6-6: Master traffic counts locations used to calculate hour-week expansion factors Master traffic counts locations used to Associated toll plazas calculate hour-week expansion factors La Apartada (M1) La Apartada Los Manguitos Purgatorio (M2) Purgatorio Cedros (M3) Cedros Mata de Caña (M5), Mata de Caña San Antero (M10) Caimanera San Onofre (M9) San Onofre Purgatorio (M2) San Carlos Source: Steer Davies Gleave The hour to week EF were not recalculated in this updated study as no new counts were conducted at any master station (24 hours, 7 days), so it was not possible to calculate these factors again. Hourly information for each day of November 2015 at La Apartada, Mata de Caña and San Onofre tolls was used to obtain expansion factors from week to month at all toll plazas as described below: La Apartada toll information: Only used for this location Mata de Caña toll information: Used to this location and for Purgatorio, Cedros, and San Antero locations. San Onofre toll information: Only used for this location Historic information from was used to analyses the seasonality of each month of the year and to obtain expansion factors from month to year for all project toll plazas. These factors were recalculated as part of the study update, to reflect data for 2016 that was available for the existing toll plazas. The data available for the two toll plazas installed on mid-2016 (Caimanera and Manguitos) was available for only few months at the time of doing this analyisis. To update the month to year EF for these toll plazas, the observed traffic volume was incorporated in the model calibration to represent the year Note that the calibration of these factors was undertaken after the assumptions regarding demand suppression and toll plaza avoidance were considered in the forecasts. More details of this are given later in this section. All the expansion factors can be found on Annex I. Percentage variation for CPI All costs and benefits were assessed using constant prices as at January This was done because this is the calibrated base year in the model. The following percentage variations on the CPI, as calculated by DANE, have been used. 2013: 1.94% Steer Davies Gleave Page 103 of 259

4 Traffic study for project Conexión Antioquia - Bolívar 2014: 3.66% 2015: 6.77% 2016: 5.75% Correction factors The objective of the opinion surveys (section 3) was to find out how demand has reacted to tariff increases and the addition of new tolls. Light vehicles are the only category that demonstrated a significant change in travel behavior. This is especially pronounced for frequent travelers and non-work related trips. Traffic reduction factors related to tariff increases were determined and are applied to each toll plaza s traffic projections over the forecast period. The correction factors are shown in the following table. Table 6-7: Price-demand factor for light vehicles Toll Price-demand factor for light vehicles La Apartada 3% Manguitos 3% Purgatorio 2% Mata de Caña 3% Caimanera 4% San Onofre 2% Source: Steer Davies Gleave Ramp-up factors Ramp-up factors were applied to traffic projections at new toll plazas (Manguitos, San Carlos and Caimanera) reflecting the uncertainty of installing a new plaza. These was a conservative assumption and was based on the observed traffic behavior in Manguitos and Caimanera after their opening in mid For ramp up factors related to the introduction of a new toll on existing routes (Manguitos and Caimanera), a one year ramp-up period was assumed, expecting 92% of total demand. We believe this should only apply for one year as users already have a good knowledge of the network and the ramp-up is only related to toll acceptancy. For ramp up factors assumed for tolls located on new roads (San Carlos), we consider this to have two different effects: first, a ramp up factor reflecting the introduction of the new toll plaza at the end of 2018, and then a factor reflecting the completion of the road works on IFU3 at the end of In this case, a four-year ramp-up period was considered. For the first two years of operation, it is expected that 84% of total demand will be captured in this station, for the following two years it is expected to reach 92% and to recover 100% of demand after year five. Toll plazas: tariffs and location The initial toll tariffs 11, expressed in pesos of the reference month (January 2014) are as shown below. They do not include the charge for the Road Safety Fund (RSF) 11 As laid down in part 4.2 Tariff Structure of the Concession Contract 016 Steer Davies Gleave Page 104 of 259

5 Traffic study for project Conexión Antioquia - Bolívar Table 6-8: Tariffs for existing project tolls (COP 2014 without RSF) TOLL CAT I CAT II CAT III CAT IV CAT V CAT VI CAT VII La Apartada $ 9,900 $ 14,700 $ 14,700 $ 14,700 $ 26,700 $ 42,800 $ 49,300 San Onofre $ 9,900 $ 14,700 $ 14,700 $ 14,700 $ 26,700 $ 42,800 $ 49,300 Los Cedros $ 9,900 $ 14,700 $ 14,700 $ 14,700 $ 26,700 $ 42,800 $ 49,300 Purgatorio $ 9,900 $ 14,700 $ 14,700 $ 14,700 $ 26,700 $ 42,800 $ 49,300 Mata de Caña $ 9,600 $ 14,300 $ 14,300 $ 14,300 $ 15,400 $ 22,400 $ 22,500 Note: These estimated do no include the Road Safety Fund RSF) Source: Concession Contract To define the toll tariffs, the following assumptions are made: All values are given in 2015 pesos, with tariffs adjusted across years using the relevant CPI. No additional increases in tariff structure were assumed, other than those laid down in the contract document. Other than these specified tolls the only tariff update from the reference month in 2015 are based on the formula specified in part 4.2 of the special part in the contract. The Concession Contract (special paragraph 4.2 indent c) and d)), establishes the formula to index the tariffs from the initial toll prices and to round the resulting values to the closest hundredth. This formula is incorporated into the tariffs in the simulation. The indent d) from special part 4.2 of the contract, specifies that the tariff includes the Road Safety Fund. The value of this is specified in the current resolution, Resolution1884 of June 17, Article 4 of this resolution specifies the value of the Road Safety fund to be $200 COP. This value will be used in 2015 values and is added to the toll tariffs in accordance with the formula in the contract. After updating the tariffs to 2015 values, adding the RSF ($200 COP from 2015) and rounding to the closest hundredth, the following tariff prices have been adopted (2015 pesos): Table 6-9: Tariff structure for project tolls and Carimagua (COP 2015 with RSF) Toll CAT I CAT II CAT III CAT IV CAT V CAT VI CAT VII Los Cedros $ 10,400 $ 15,400 $ 15,400 $ 15,400 $ 27,700 $ 44,300 $ 51,100 Purgatorio $ 10,400 $ 15,400 $ 15,400 $ 15,400 $ 27,700 $ 44,300 $ 51,100 La Apartada $ 10,400 $ 15,400 $ 15,400 $ 15,400 $ 27,700 $ 44,300 $ 51,100 Los Manguitos $ 10,400 $ 15,400 $ 15,400 $ 15,400 $ 27,700 $ 44,300 $ 51,100 San Carlos $ 10,100 $ 15,000 $ 15,000 $ 15,000 $ 16,100 $ 23,300 $ 23,400 Mata de Caña $ 10,100 $ 15,000 $ 15,000 $ 15,000 $ 16,100 $ 23,300 $ 23,400 La Caimanera $ 10,100 $ 15,000 $ 15,000 $ 15,000 $ 16,100 $ 23,300 $ 23,400 San Onofre $ 10,400 $ 15,400 $ 15,400 $ 15,400 $ 27,700 $ 44,300 $ 51,100 Carimagua $ 10,400 $ 15,400 $ 15,400 $ 15,400 $ 27,700 $ 44,300 $ 51,100 Source: Concession Contract After the Concession went into effect (October 2015), modifying resolutions were issued to the tariffs for some of the project tolls and others in the study area. The tariffs changes applied at: Los Manguitos, Mata de Caña, San Onofre, Caimanera and Carimagua. Steer Davies Gleave Page 105 of 259

6 Traffic study for project Conexión Antioquia - Bolívar The following subsections present the situation for each station and the tariffs considered for the model (expressed in 2015 COP) in accordance with the issued resolutions by the Ministry of Transport up to the date of this study. Toll La Apartada Changes to tariff structure: There have been no modifying resolutions to the tariff for this toll. The Concession Contract does not specify a different increase under numeral 4.2 section e) of the special part of the contract. Location: According to the conditions of the Concession Contract, this toll will not be moved during the concession period. Start of operation: This toll belongs to the concession at the moment of the starting act (October 2015). The tariff increases established in the contract are considered from January 2016 on. Toll Los Manguitos Changes to tariff structure: Resolution 2820 of 11 July 2016 establishes new tariffs for categories I, IV and VII, to 2016 COP with RSF fee. These are shown in the following table: Table 6-10: Special tariffs for Los Manguitos (COP 2016 with RSF) Categories Special Categories COP 2016 (Con FSV) Automobiles, SUVs and pickup trucks E IE $5,700 5 axles trucks E VIE $25,000 6 axle trucks E VIIE $25,000 Source: Resolution 2820 of 11 July 2016 The new tariffs apply in the following cases: o The new category IE tariffs apply only to vehicles having an Electronic Identification Card from 13 public transport companies and Buenavista taxi. This equates to 345 vehicles in category I. o The impact of these tariffs is not considered in the projections since it only applies to 345 vehicles, and its effect is expected to be marginal. o For categories VIE and VIIE (Large trucks), these tariffs will apply to all vehicles using this station, until at least two of the following cases occur: More than 2.5 years have passed since the resolution San Carlos toll starts operation More than 50% of the cost of IFU3 has been invested and there is free circulation in the IFU. The construction calendar indicates that these three cases are expected to occur in January Therefore, our forecasts assume that from January 2019 the toll at Los Manguitos returns to the tariff defined in the Concession Contract. Below is the assumed tariff structure for this location: Table 6-11: Tariff structure for Los Manguitos (COP 2015 with RSF) Category Modelling scenarios Tariffs in COP 2015 with RSF forward I $ 10,400 $ 10,400 $ 10,400 II $ 15,400 $ 15,400 $ 15,400 III $ 15,400 $ 15,400 $ 15,400 IV $ 15,400 $ 15,400 $ 15,400 V $ 27,700 $ 27,700 $ 27,700 Steer Davies Gleave Page 106 of 259

7 Traffic study for project Conexión Antioquia - Bolívar Category Modelling scenarios Tariffs in COP 2015 with RSF forward VI $ 23,600 $ 23,600 $ 44,300 VII $ 23,600 $ 23,600 $ 51,100 Source: Steer Davies Gleave based on the resolution 2820 July 11, 2016 Resolutions 2820 of 11 July 2016 and 5711 of 26 December 2016 establishes a contractual location change for this station as follows: o Location established in the concession contract: PR o Location established in resolution : to PR o Location established in resolution : This plaza is currently located at PR but it is going to be relocated for a permanent basis topr once the IFU1 is completed in early In accordance with the Concession Contract (3.6 indent f), this toll was installed in June 24, Thus, the model starts with January 2016 values and the tariff structure is as shown in previous table. Tolls Cedros and Purgatorio Changes to tariff structure: There is no resolution to modify these tolls. The tariff structure is as in Table 6-8 Location: The special part of the Concession Contract (3.6 paragraph d), establishes that these tolls will be relocated in June 2017, when the IFU to which they belong is transferred to the Ruta del Mar concession from the current operator, Transversal de las Américas. o Toll Purgatorio relocation in July 2017: from PR to PR o Toll Los Cedros relocation in July 2017: from PR to PR These tolls will be charged with the tariff structure indicated in Table 6-9 from July 2017, when they are transferred to the Concession. For modelling scenarios, it is assumed that these tolls and their respective functional units are opened to the public in 2018, considering that it could take a few months to relocate them. These tolls have not been relocated as of the date of the preparation of this report and the traffic forecasted in this report is highly dependent on the final location of the toll plaza. Toll Mata de Caña Changes to tariff structure: As with Los Manguitos, Resolution 2820 of July 11, 2016 establishes special rates for three public transport companies that operate in category II and whose fleet sum 320 vehicles. The special tariff for these companies is $12,000 COP (2016), including RSF fee. As at Los Manguitos, the impact of these tariffs is not considered in the projections since they only apply to a few vehicles and is not material to the model. Location: This toll will not be moved during the concession period. Start of operation: The Mata de Caña station is part of the concession from the time of the signing of the initiation act (October 2015). In the model, it is considered that the tariffs are increased in January 2016, in accordance with Table 5-9. Toll Caimanera Changes in tariff structure: Resolution 3119 of July 29, 2016 regulates special rates for all vehicle categories provided they meet the following requirements: o Public transport vehicles (passengers and freight) that provide the service on the Tolú - Coveñas route and vice versa which use the toll plaza at least 30 days in a month. Steer Davies Gleave Page 107 of 259

8 Traffic study for project Conexión Antioquia - Bolívar o Private vehicles owned by people living in Tolú or Coveñas which use the toll plaza at least 30 days a month. These special tariffs are permanent for the concession period, apply for all vehicles that meet the criteria described before, and have the following values in 2015 pesos with RSF: Table 6-12: Tariff structure for Caimanera (COP 2015 with RSF) Categories Type Tariffs COP 2015 with RSF Automobiles, SUVs and pickup trucks I $10,100 Automobiles, SUVs and pickup trucks E IE $700 Buses II $15,000 Buses E IIE $700 Small 2 axle trucks III $15,000 Small 2 axle trucks E IIIE $700 Large 2 axle trucks IV $15,000 Large 2 axle trucks E IVE $700 3 and 4 axle trucks V $16,100 3 and 4 axle trucks E VE $700 5 axle trucks VI $23,300 5 axle trucks E VIE $700 6 axle trucks VII $23,400 6 axle trucks E VIIE $700 Source: Resolution 3119 July 29, 2016 The special rates indicated in the table apply only to vehicles that use the Caimanera Toll for at least 30 days a month, and only travel between Tolú and Coveñas. In the network model, it is considered that these vehicles pass through the plaza and that as indicated, the Concession controls the bypass routes. Still the number of vehicles to which the special tolls apply is low and therefore these special rates are not included in the network model vehicles. Toll San Onofre Location: Resolution 3119 of July 29, 2016 also establishes a change in the location of this station as follows: o Location established in the Concession Contract: PR o Current Location and established by Resolution 3119 of July 29, 2016: PR Changes in tariff structure: Resolution 2820 of July 11, 2016 establishes special tariffs that began to be charged as of the date of issue for categories VI and VII. These are presented in the following table: Table 6-13: Special tariffs for San Onofre (COP 2016 with RSF) Categories Special categories COP 2016 (with RSF) 5 axle trucks E VIE $20,800 6 axle trucks E VIIE $23,900 Source: Resolution 2820, 2016 Steer Davies Gleave Page 108 of 259

9 Traffic study for project Conexión Antioquia - Bolívar According to resolution these tariffs apply in the following cases: All the vehicles categories VI and VII using this station For 2.5 years from the resolution issuance (11 July 2016) Until San Onofre toll starts operation Until more than 50% of the cost of IFU3 has been invested and there is free circulation in the IFU Based on the schedule of construction, these three points are expected to occur in January Furthermore, Resolution 2820 established the following tariff structure. Table 6-14: Tariff structure for San Onofre (COP 2015 with RSF) Modelling scenarios Tariffs in COP 2015 with RSF Category forward I $10,400 $10,400 $10,400 II $15,400 $15,400 $15,400 III $15,400 $15,400 $15,400 IV $15,400 $15,400 $15,400 V $ 27,700 $ 27,700 $ 27,700 VI $ 19,700 $ 19,700 $44,300 VII $22,600 $22,600 $ 51,100 Source: Resolution 2820, 2016 Location: the agreement established between ANI and Ruta del Sol, this toll will not be relocated for the concession period. Start of operation: This toll is part of the concession from the start (October 2015). In the model, the tariffs are increased in January 2016, in accordance with the table above. Toll Carimagua Tariff structure: Even though this toll is not part of the concession, the Concession Contract established that once the initiation act was signed (October 2015), the tariff structure of this toll plaza will be as per Table 6-9. In addition, resolution 2819 of July 2016 suspends the toll for categories VI and VII in Carimagua until one of the following conditions is met. o 2.5 years from the resolution issuance (11 July 2016) o San Onofre toll starts operation o More than 50% of the cost of IFU3 has been invested and there is free circulation in the IFU It is expected that at least one of these criteria will occur by January of Consequently, the Carimagua toll is modelled using the tariff structure outlined as follows: Steer Davies Gleave Page 109 of 259

10 Traffic study for project Conexión Antioquia - Bolívar Table 6-15: Tariff structure for Carimagua (COP 2015 with RSF) Modelling scenarios Tariffs in COP 2015 with RSF Category forward I $10,400 $10,400 $10,400 II $15,400 $15,400 $15,400 III $15,400 $15,400 $15,400 IV $15,400 $15,400 $15,400 V $ 27,700 $ 27,700 $ 27,700 VI $0 $0 $44,300 VII $0 $0 $ 51,100 Source: Concession Contract and resolution 2819, 2016 Location: The Concession Contract establishes that the toll plaza is moved to PR on the road between Planeta Rica and La Y and this is the location which has been modelled. Steer Davies Gleave Page 110 of 259

11 Traffic study for project Conexión Antioquia - Bolívar Future Projects Changes in the wider highway network will also impact on traffic using the concession. The following table lists the projects planned to be built within the study area during the life of the concession. The information presented here corresponds to the revision of contracts, technical annexes, and public domain tender documents available at the date of preparation of this document. Table 6-16: Future projects No. Project Description Operation base year Project status Information source Ruta del Sol sector km dual carriageway between Puerto Salgar and San Roque. It will have a dual carriageway on each of its seven segments. Full operation in 2021 Contracted ongoing SECOP web page ANI web page 2 Ruta del Sol sector km project with three dual carriageway corridors. Full operation in 2021 Contracted ongoing SECOP web page ANI web page 3 Autopista Puerta de Hierro - Palmar de Varela - Carreto - Cruz del Viso 202 km project, covering 14 municipalities of Sucre, Bolívar and Atlántico. It considers the construction, operation and maintenance of the existing roads between Puerta de Hierro and Palmar de Varela, and the construction of some connectioning roads to nearby municipalities. Full operation in 2020 Contracted SECOP web page ANI web page 4 Concesión Vial Córdoba - Sucre Autopistas de la Sabana S.A. is the concessionaire in charge of this project which includes the construction of 120 km of double carriageway and rehabilitation of 260 km. The following sections will see construction of second carriageways: Sincelejo-Corozal, Sincelejo-Sampués, La Ye- Sahagún, Dual carriageway parallel to Montería beltway, Cereté - La Ye, Montería - Cereté, Sincelejo - Tolúviejo. Sampués - Sincelejo and Sincelejo - Tolúviejo Started in December 2016 Concession ends June 30, 2024 (Addendum7) Contracted ongoing SECOP web page ANI web page 5 Corridor Cartagena - Barranquilla - Ruta Caribe This project crosses the municipalities of Cartagena, Santa Catalina, Tubará, Juan de Acosta, Piojo, Puerto Colombia and Barranquilla, and is approximately 90 km long. Full operation in 2020 Contracted ongoing SECOP web page ANI web page Steer Davies Gleave Page 111 of 259

12 Traffic study for project Conexión Antioquia - Bolívar No. Project Description Operation base year Project status Information source 6 Autopista Mar 1 An extension of 176 km providing connection between Medellín and Bolombolo with Santa Fe de Antioquia. The sections Medellín - Bolombolo - Santa Fe de Antioquia will be improved. The interventions include: Improvement of existing roadway, construction of a second carriageway on the Medellín-Santa Fé de Antioquia section, the second phase of the West Tunnel Fernando Gómez Martínez (a new parallel tunnel of equal length to the existing one 4,7 km), the repairs of the Bolombolo-Santa Fe section and the operation of the road from Santa Fe to Cañasgordas 2021 Contracted SECOP web page ANI web page 7 Autopista Mar 2 Extension of 246 km providing connection between Cañasgordas and Urabá Antioqueño reaching the point known as El Tigre in the municipality of Chigorodó. Interventions include: Single carriageway road in all sections, with varying design speeds per section: Canasgordas -Dabeiba, (80km / h), Dabeiba-El Tigre (70km / h), El Tigre-Necocli (as current) 2021 Contracted Financial closure SECOP web page ANI web page 8 Toyo Tunnel Connecting the Concession Autopista al Mar 1 and 2. The tunnel will be located between the municipalities of Santa Fe de Antioquia and Cañasgordas. This section will be 39 kilometers and includes the construction of several tunnels, reducing the current length by 23 km. This work doubles the current design speed of the road. Full operation in 2025 Contracted SECOP web page 9 Transversal de Las Américas 714 km in eleven sections. The relevant section to Ruta al Mar project are as follows: -El Tigre - Chigorodó: Repair of single bidirectional carriageway in two lanes. Currently operating. -Chigorodó - El Tres: Construction of second carriageway with an expected design speed of 80 km/h. -El Tres - Puerto Rey: Repair, improve and build relief roads. Bidirectional single carriageway road with two lanes. Design speed 80 km/h along most of the route. Currently operating according to the press release by ANI 12 Contracted ongoing SECOP web page ANI web page 10 Autopista Conexión Norte Connecting the municipalities of Remedios, Zaragoza and Caucasia; connects the south west and center west of the country. This project is a single carriageway with an approximate length of 145 km, starting in the municipality of Remedios in Antioquia Contracted SECOP web page ANI web page 12 Steer Davies Gleave Page 112 of 259

13 Traffic study for project Conexión Antioquia - Bolívar No. Project Description 11 Autopista al Rio Magdalena 1 (Vía del Nus) Starts in Bello passing through Hatillo, Portecito and Cisneros, to Alto de Dolores, including Hatillo - Don Matías section. 93km of new road construction and the operation and maintenance of another 156km. The project is divided into the following sections: -Pradera - Porcesito: new road construction -Porcesito - San José del Nus: improvement and double carriageway, tunnel construction in double carriageway, -San José del Nus - Alto de Dolores: construction of a new carriageway 12 Autopista al Rio Magdalena km long and has the following segments and interventions: -Remedios - Vegachí: Single carriageway construction -Vegachí - Alto de Dolores: Single carriageway construction -Alto de Dolores - Puerto Berrío: Improvement and repairs -Puerto Berrío - Conexión Ruta del Sol: improvement and construction 13 Concesión Vial Autopistas del Caribe 253 km between Cruz del Viso and Barranquilla in the following IFU: -IFU0 Operation and maintenance of the existing roads of the Concesión Ruta Caribe -IFU1 Construction of second carriageway on a 6,37 km stretch between Cruz del Viso and Gambote and installation of a new Toll booth. -IFU2 Gambote - Corrales - Gamboa - Cartagena: -IFU3 Construction of second road plus Toll station between Bayunca and Clemencia (7,4 km). Construction of single carriageway Bayunca relief road and roundabout at intersection with 90A -IFU4 Construction of second carriageway between Clemencia and Santa Catalina (6,6 km). Construction of a single carriageway Clemencia relief road -IFU5 Construction of second carriageway between Santa Catalina and Luruaco (12,1 km). Construction of a single carriageway relief road for Pendales -IFU6 Construction of a single carriageway relief road to Luruaco. Construction of access to the relief road Luruaco. -IFU7 Second road Luruaco - Sabanalarga (23,1 km): Construction of a new toll booth between the towns of Arroyo de Piedra and Molinero. -IFU8 Construction of a single carriageway Sabanalarga relief road. -IFU9 Campeche - Road crossing to Caracolí: Construction of a new toll booth on the new road near Galapa toll. Second carriageway construction between Campeche and the road junction to Caracolí (17,3 km). Construction of a single carriageway relief road Baranoa. Bridge construction at the crossing of the road Polo Nuevo - Baranoa with Baranoa relief road. Operation base year Project begins operation at different years: Full operation of IFU1, IFU2, IFU3, IFU9 in Full operation for UF4, UF5, UF6, UF7 and UF8 in 2025 Project status Contracted Preconstruction Tender in process Information source SECOP web page ANI web page SECOP web page ANI web page SECOP atos.gov.co/consu ltas/detalleproces o.do?numconstan cia= Steer Davies Gleave Page 113 of 259

14 Traffic study for project Conexión Antioquia - Bolívar No. Project Description Operation base year Project status 14 New IP Córdoba - Sucre Design, construction, operation and maintenance of a second road for Ciénaga de Oro - La Yé; Sahagún - Chinú; Chinú - Sampues. Design, construction, operation and maintenance of a relief road in the municipality of Sahagún; a relief road in the municipality of Chinú; and relief road in the municipality of Samsepues. Design, rehabilitation, improvement, operation and maintenance of existing facilities at Planeta Rica - La Ye and La Ye - Sincelejo. Full operation in 2025 Feasibility study since 17/10/ Desarrollo vial bajo Sinú (Lorica - Chinú) Private initiative without public resources that considers the sustainable maintenance of Lorica - Chinú corridor. Full operation in 2023 Feasibility study since 12/05/ PPP of private initiative for Corridor Planeta Rica - La Ye, in Córdoba Private initiative without public resources that considers the sustainable maintenance of the road Planeta Rica - La Ye, Ruta 25, segment This project only implies the maintenance of the road. It is considered that this segment will be in good conditions constantly until Feasibility study since 05/09/2016 Source: Steer Davies Gleave based on contracts from each project in SECOP and the ANI (National Infrastructure Agency) website Information source RUAPP - DNP 13 RUAPP - DNP RUAPP - DNP Steer Davies Gleave Page 114 of 259

15 Traffic study for project Conexión Antioquia - Bolívar Key dates for project execution Section 5.2 of the Concession Contract defines the Schedule of construction stating the completion dates of each IFU. These are assumed in the network model: Table 6-17: Key dates for project execution IFU Sector Months from initiation act for construction Starting operation Model starting year IFU1 Caucasia - Planeta Rica Jan IFU2 Cereté - Lorica Jan IFU3 El 15 - San Carlos - Cereté Dec IFU4 Planeta Rica - Montería 30-Jul IFU5 Puerto Rey - Montería, Santa Lucía - San Pelayo 30-Jul IFU6 Cereté - Tolú Dec IFU7 Relief roads Lorica and Coveñas + Tolú - Pueblito Jan IFU8 Tolú - Tolúviejo - San Onofre - Cruz del Viso Jan Source: Steer Davies Gleave based on the Concession Contract Modelling years The modelling years for the base scenarios have been agreed based on the following: The construction schedule for the project and opening years for each IFU. Opening years of projects listed in Table Years when the tariffs or toll locations change within the Study Area. Below are the assumed model years and the key considerations for each one: Year 2015 Base year, calibration scenario Year 2016 Calibration and update year for growth and expansion models Year 2019 Start of operation of IFU1 and missing segment of IFU3 (Improvement of existing road between San Carlos and double carriageway Cereté - Ciénaga de Oro) Los Manguitos, San Onofre and Carimagua tolls change their tariff structure in line with Table 6-9. Full operation for double carriageway Chigorodó - El Tres of Transversal de las Américas concession. Steer Davies Gleave Page 115 of 259

16 Traffic study for project Conexión Antioquia - Bolívar Year 2020 Year 2021 Year 2023 Full operation for Autopista Puerta de Hierro - Palmar de Varela - Carreto - Cruz del Viso and Corridor Cartagena - Barranquilla The following IFU begin operation: o IFU2: Cereté - Lorica o IFU6: Cereté - Tolú o IFU7: Variantes Lorica and Coveñas and Tolú - Pueblito o IFU8: Tolú - Tolúviejo - San Onofre - Cruz del Viso The following projects begin operation: o Autopista Mar 1 o Autopista Mar 2 o Autopista Conexión Norte o Autopista al Rio Magdalena 1 (Vía del Nus) o Ruta del Sol sector 2 o Ruta del Sol sector 3 Current Córdoba - Sucre (project 4 on Table 6-16) concession ends. The following projects begin operation: o Desarrollo vial bajo Sinú (Loríca - Chinú). o The following Integral functional units for Concesión Vial Autopistas del Caribe (Project number 15 of the table): IFU1, IFU2, IFU3, IFU9 Year 2025 The following projects begin operation: o IP Córdoba - Sucre (IP 14) o Toyo Tunnel Concesión Vial Autopistas del Caribe commences following IFUs (Project number 15 on Table 6-16): IFU4, IFU5, IFU6, IFU7, IFU8 Year 2030 Medium term scenario Year 2035 Long term scenario From 2040 a constant traffic growth rate of 2.0% is assumed. This is appropriate given the uncertainty of longterm forecasts. Projections will be provided on this basis up to 2055 as required by the Concession financial team. The following images show the project applicable for each future year scenario as described above. Steer Davies Gleave Page 116 of 259

17 Traffic study for project Conexión Antioquia - Bolívar Figure 6.7: Road network configuration in the study area for future years Year 2019 Year 2020 Steer Davies Gleave Page 117 of 259

18 Traffic study for project Conexión Antioquia - Bolívar Year 2021 Year 2023 Steer Davies Gleave Page 118 of 259

19 Year 2025 Source: Steer Davies Gleave Traffic study for project Conexión Antioquia - Bolívar Steer Davies Gleave Page 119 of 259

20 Traffic study for project Conexión Antioquia - Bolívar 7. DETAILED ANALYSIS OF TRANSPORT COSTS CURRENT PROJECT AND TRAFFIC CHARACTERISTICS The project covers a total length of 492 kilometers, of which 112 kilometers are new roads, 226 kilometers are existing roads that will be improved, and 154 kilometers will be operated and maintained after being transferred from the current Concesión Transversal de Las Américas. The concession is divided into eight IFUs shown below. Existing roads are shown using solid lines while roads to be constructed are identified with dashed lines. Between Cereté and Lorica the concessionaire will build a second carriageway. Figure 7.1: Outline and Functional Units of the project Source: Steer Davies Gleave Steer Davies Gleave Page 120 of 259

21 Traffic study for project Conexión Antioquia - Bolívar The project corridor between Caucasia and Cartagena is currently incomplete. The current route is longer than the main competing route due to the lack of a direct connection between Tolú and San Onofre. In addition, this corridor is characterized by numerous urban sections through important towns along the route such as Planeta Rica, Monteria, Lorica, Coveñas, Tolú. Thus, the market for this route is primarily short and medium distance trips. It is anticipated that once the project is built, this traffic will continue to use these roads as there are no other comparable travel alternatives. Most of the long-distance traffic between the department of Antioquia and the north of the country (in particular Cartagena) currently uses the competing corridor Planeta Rica - La Y - Sincelejo - Carmen de Bolívar - Cruz del Viso. This is shown on the map above as the Current long distance route. The section between La Y and Puerta del Hierro forms part of the Córdoba - Sucre Concession that began operating in 2008 and is currently expanding some stretches to dual carriageway. The analysis presented below compares the project and the alternatives. As short and medium distance trips are not expected to change routing, it focuses on long distance traffic. ROUTE COMPARISONS The main objective of the Concesión Ruta al Mar is to capture the long-distance traffic between the northern part of the coffee axis, Medellín, the Caribbean Coast area of the departments of Córdoba and Sucre, and its connection with Cartagena. This long-distance traffic is heavily comprised of trucks (90%). However, the competing corridors are also under plans for new projects and road concessions that will impact the times and costs of travel in the study area. The following figure shows the various schemes on the competing corridors, including the years of commencement. The improvements and delivery dates of the competitive schemes will impact the capture of long distance traffic across the life of the concession. Three periods were reviewed to understand the impact of the different schemes on the capture rate of project traffic : Ruta al Mar is incomplete and competition is with the Córdoba - Sucre concession; : Ruta al Mar is complete and competitive roads are the Córdoba - Sucre concession and the Puerta de Hierro - Palmar de Varela road; 2025-end of the concession: Ruta al Mar is complete and competitive roads are the Planeta Rica - Sampués and Puerta de Hierro - Palmar de Varela concessions. The aspects of the project corridor are compared with the competition for each of the periods considered. The traffic capture rates for the project road is estimated to sense check the results of the network model. This analysis focuses on cargo traffic as trucks represent the clear majority of long distance traffic. Steer Davies Gleave Page 121 of 259

22 Traffic study for project Conexión Antioquia - Bolívar Figure 7.2: Project Competitive corridors Source: Steer Davies Gleave : Ruta al Mar is incomplete and its competition is the concession Córdoba - Sucre In this period, 80% of long distance traffic is large trucks. 74% of light vehicles travelling between the department of Antioquia and Cartagena will continue to use the competing route between Planeta Rica, Sincelejo and San Onofre (Route 3 of the following map) for the following reasons: The competing road is shorter that the project road by 40 km The competing road is faster by 40 to 50 minutes The competing road has lower toll costs compared to the project road and route 2: o For light vehicles, it costs $3,700 COP less than in the project o For small trucks, it costs $9,900 COP less o For large trucks, it costs $12,000 COP less To illustrate the points above, the following figure shows the 2015 distances, travel times, and fares of light vehicles (LV), small trucks (ST) and large trucks (LT) over the three routes between Caucasia and Cartagena. Steer Davies Gleave Page 122 of 259

23 Traffic study for project Conexión Antioquia - Bolívar Figure 7.3: Route alternatives (Caucasia - Cartagena) in 2015 Source: Steer Davies Gleave : Ruta al Mar is complete and its competing roads are the concession Córdoba - Sucre and Puerta de Hierro - Palmar de Varela Once the new roads (IFU2, IFU3 and IFU7), are completed, the project (Route 1) is expected to be 20 to 30 minutes faster than route 2, and have a similar travel time to route 3. Although the distances between Cartagena and Medellín are very similar for either route 1 or route 2, the difference in travel times are linked to the number of urban areas passed through. Route 2 passes through several urban areas while route 1 has several relief roads avoiding urban areas. The project toll tariffs are greater than in the two alternate routes considered. Light vehicles will pay about $16,000 COP more than route 2 and about $13,000 COP more than route 3. A similar situation occurs for small trucks where the difference is about $22,000 COP. Large trucks pay about $20,000 COP less for Route 2 and Steer Davies Gleave Page 123 of 259

24 Traffic study for project Conexión Antioquia - Bolívar about the same for Route 3. This explains the capture rate for large trucks between 2021 and 2024, which rises to more than 35% of all heavy truck traffic travelling between the department of Antioquia and Cartagena. The capture rate for light vehicle and small truck traffic will remain low (7% and 2% respectively for trips between Caucasia and Cartagena). Figure 7.4: Route alternatives (Caucasia - Cartagena) in 2021 Source: Steer Davies Gleave 2025-End of concession: Ruta al Mar is complete and its competition roads are the concessions Planeta Rica - Sampués and Puerta de Hierro - Palmar de Varela. It is assumed that the Córdoba - Sucre concession begins operation in This infrastructure reduces the travel times and increases the toll costs of Routes 2 and 3. This is due to the introduction of a new toll plaza to the south of Sincelejo and to the increase of tariffs at the La Esperanza toll (north of Sincelejo). Steer Davies Gleave Page 124 of 259

25 Traffic study for project Conexión Antioquia - Bolívar The rate increases between 2021 and 2024 will be approximately $10,400 COP for light vehicles, $15,000 COP for small trucks and $40,900 COP for large trucks on route 2; and $14,400 COP for light vehicles, $23,000 COP for small trucks and $62,000 COP for large trucks on route 3. These changes in times and costs mean that the project (Route 1) will become a more attractive route. Large trucks will save almost $20,000 COP compared with Route 2 and $40,000 COP compared with Route 3. Between Caucasia and Cartagena, this leads to an increase in large truck traffic from 37% to more than 90%. Small truck traffic is estimated to rise from 2% to 12%, and light vehicles from 7% to 44%. This reflects the smaller difference in tariffs. Figure 7.5: Route alternatives (Caucasia - Cartagena) in 2025 Source: Steer Davies Gleave Steer Davies Gleave Page 125 of 259

26 Traffic study for project Conexión Antioquia - Bolívar CONCLUSIONS The key points from this section are: Existing segments of the Concesión Ruta al Mar, are currently used mainly by individuals travelling short and medium distances between municipalities. It is expected that this traffic will continue using the road in the future; The main competing routes to the project are: o The Sahagún - Sincelejo route and Carmen de Bolívar (Route 2), which currently serves short and medium distance traffic between the municipalities of Sahagún, Chinú, Sincelejo, Carmen de Bolívar, Carreto, etc. Long distance traffic between the department of Antioquia and Barranquilla and part of the long-distance traffic between the department of Antioquia and Cartagena are also served by this route; o The Sahagún - Sincelejo and San Onofre Route (Route 3), which currently serves short and medium distance traffic between the municipalities of Sahagún, Chinú, Sincelejo, Tolúviejo, San Onofre, etc. Most long-distance traffic between the department of Antioquia and Cartagena also use this route. Following the completion of the Project's IFUs in 2021, the Ruta al Mar Concession is expected to capture part of the long-distance traffic between the department of Antioquia and Cartagena which currently use route 2 or 3. The demand for Ruta al Mar is expected to consist mostly of trucks (approximately 90%). Between 2021 to 2024 the number of light vehicles and small trucks using Ruta al Mar is expected to represent only a limited share, approximately 10%, of demand between the department of Antioquia and Cartagena. This is due to the relatively higher tariffs compared with alternatives. The capture rate for large trucks is forecast to be around 37% of the total market. From 2025 onwards, the traffic capture of light vehicles, small trucks and large trucks is expected to rise respectively to 44%, 20% and 92% due to the tariff increases of Route 2 and 3, generated by the entry into operation of the Córdoba - Sucre concession. Light vehicle traffic will be much less affected by the competing corridors, as the long-distance light vehicle trips comprise a relatively small share of the total light vehicle demand. Steer Davies Gleave Page 126 of 259

27 Traffic study for project Conexión Antioquia - Bolívar 8. TRANSPORT FLOW ANALYSIS This section summarizes the main results of the baseline traffic projections for the Ruta al Mar Concession. The results are presented for both the current and future tolls specified as part of the Concession. The forecasts of traffic and revenue contained in this report were produced in March 2017 and as such inputs, assumptions and data employed within the forecasting process are derived from information available as of March This report is dated November 2017 and it, and the forecasts contained within it, do not take into account any additional or amended information that has become available or known since March In that regard we are aware of the following information that could, if incorporated into our forecasting process, lead to changes to the forecasts of traffic and revenue that are currently contained within this report: 2. The downwards revision for 2017 and 2018 of GDP growth forecasts produced by the Government of Colombia and other financial institutions. 3. The previously planned relocation of some toll plazas (Purgatorio and Cedros) has not yet occurred but this relocation was accounted for within our forecasting process and within the results presented in this report starting for the full calendar year We have been provided with evidence confirming an increasing evasion pattern in some plazas, particularly La Apartada and Caimanera but not limited to those. The traffic forecasts included in this report do not account for toll evasion and caution should be made in regard to its consideration as the number to be used to produce revenue forecasts. 5. We note the increased likelihood of the development of a port in Urabá. This observation is not accounted for in our base case scenario but informed in one of our contrast scenarios. TRAFFIC COMPOSITION The traffic forecasts at each of the concession toll plazas are presented in terms of AADT, characterized by the three components below: Existing traffic current traffic in the segments where toll plazas are or will be located. Diverted or attracted traffic increase or decrease in traffic at each toll plaza caused by network or tariff changes in the study area. The changes are either by the concession or other schemes. Demand growth additional traffic anticipated due to the increase in the demand from light and heavy vehicles in the study area. The methodology used to forecast this is described in Chapter 5. Note that: For each toll plaza, diverted traffic was calculated by comparing the forecast traffic in two scenarios: Full 2025 network with 2025 demand; and Full 2025 demand with the 2015 base network. For each toll plaza, demand growth was calculated by comparing two scenarios: Full 2025 network with 2025 demand; and Full 2025 network with current 2015 demand. The demand study assumes no induced demand. The segmentation of the 2025 traffic projection includes some approximations on which trips are diverted and which are due to demand growth. Demand growth changes travel times in the network and consequently influences the capture rates for the Ruta al Mar Concession. We consider that these approximations are small enough to not alter the conclusions of the analysis. Steer Davies Gleave Page 127 of 259

28 Traffic study for project Conexión Antioquia - Bolívar The following figure shows the composition of total traffic across all the toll plazas in The percentages of existing traffic, deviated traffic, and demand growth were calculated by dividing the estimated traffic for each component by the total traffic for each toll plaza. This year has been chosen as it is the first year of full operation of Ruta al Mar. Figure 8.1: Total traffic by component by toll plazas in the concession in 2025 with full operation Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. From the graph, above, we note the points below to provide some context on the wider traffic forecasts and analysis. La Apartada and Los Manguitos (on the Caucasia - Montería corridor): The traffic projected is mainly existing traffic. This is traffic travelling between Antioquia, the center west of the country, and the Atlantic coast (mainly the area of Monteria, Cartagena, and Barranquilla). These trips should pass through the toll plazas of La Apartada and Manguitos independently if they are going to use Ruta al Mar or the competing route via Planeta Rica - Sahagún and Carmen de Bolívar. The opening of the IP Planeta Rica - La Y - Sincelejo route does not significantly affect the transactions in these plazas, as these trips such as Caucasia - Cartagena are captured regardless of route. The transactions at these locations are forecast to lose between 3% and 4% traffic, as a result of the opening of alternate routes between the central western area of the country and the Caribbean coast through Urabá, when projects such as 1, and Tunel del Toyo are completed. Cedros and Purgatorio Most of the expected future traffic already uses the Ruta al Mar route (Purgatorio 78% and Cedros 73%). Steer Davies Gleave Page 128 of 259

29 Traffic study for project Conexión Antioquia - Bolívar The Purgatorio toll loses approximately 4% traffic due to the opening of alternate routes to reach Montería from Medellín via Urabá. San Carlos This is a new toll plaza and so does not have any existing traffic. Future demand is a combination of diverted traffic, most of which currently passes through Carimagua, and forecast annual growth. Mata de Caña, Caimanera, and San Onofre These plazas are heavily influenced by the opening of the IP Córdoba - Sucre segment as they are in places where long distance trips have an alternative to Ruta al Mar via Planeta Rica - Sahagún and Carmen de Bolívar. In 2025 the Ruta al Mar route becomes much more attractive (see previous chapter) due to the increase in competitive tolls. Thus, it is forecast that these locations capture the following shares of traffic: o Mata de Caña: 27% o Caimanera: 33% o San Onofre: 21% Growth (independent of trip diversion) through to 2025 represents between 19% and 24% of traffic. BASE SCENARIO TRAFFIC PROJECTIONS This section presents the analysis of the traffic projections for the Concession. For each toll station the AADT traffic and annual growth are presented for the period La Apartada The key results are summarized in the following points: Growth varies between 3% and 4% per year until 2020 largely due to annual growth. This is because all long-distance trips pass through this location regardless of the route taken north of Planeta Rica. In the year 2025, following the opening of alternative routes from Medellín to the Caribbean via Urabá a lower growth rate of approximately 2% is observed. These routes include Tunel del Toyo and the Mar 1 and Mar 2 concessions. These routes provide an alternative to the Medellín - Caucasia - Cartagena route. From 2025 growth in total AADT of around 3% per year is observed up to Most of this growth is due to annual traffic growth forecast based on expected GDP growth. Steer Davies Gleave Page 129 of 259

30 Traffic study for project Conexión Antioquia - Bolívar Figure 8.2: Base scenario AADT traffic projections La Apartada Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 130 of 259

31 Traffic study for project Conexión Antioquia - Bolívar Steer Davies Gleave Page 131 of 259 Table 8-1: Base scenario AADT traffic projections La Apartada Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas.

32 Traffic study for project Conexión Antioquia - Bolívar Los Manguitos The behavior at this toll is very like the La Apartada toll. Both are between Caucasia and Planeta Rica, a relatively homogeneous sector, which doesn t have any major traffic attractors or generators. There are no roads which can divert or inject traffic to Ruta al Mar. This means that the captured demand is mostly long distance trips travelling between the center west of the country and the Caribbean Coast regardless of whether Rutal al Mar or an alternative route is used. Figure 8.3: Base scenario AADT traffic projections Los Manguitos Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 132 of 259

33 Traffic study for project Conexión Antioquia - Bolívar Steer Davies Gleave Page 133 of 259 Table 8-2: Base scenario AADT traffic projections Los Manguitos Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas.

34 Traffic study for project Conexión Antioquia - Bolívar Purgatorio Purgatorio Toll Plaza sees an important flow growth between 2017 and 2018, following the relocation that is expected to occur at the end of 2017 closer to Monteria. The projections listed in the previous table, considers that this station is installed in the new location from the beginning of 2018 (January 1rst). Fieldwork results presented in Chapter 3 show that growth can be expected as a result of the plazas future location. The weekly average daily traffic (AWDT) found in November 2015 in the plazas current location is vehicles (M2 master station). The count carried out in the last week of January 2017 (M14 master station) on is about 70% higher (5.176 vehicles). Even though these two values cannot be directly compared because they were measured with a year of difference and in different months, they do indicate how different the traffic can be with the new plaza location. A high traffic portion (about 70%) is made up of light vehicles (Category I) due to its proximity to population centers. Due to the reduction in demand anticipated by the installation of a new toll plaza, a demand correction factor is applied to the traffic projections to reflect the future toll station location. According to the opinion surveys results described in chapter 4, about 3% of light vehicles drivers indicated they would modify their current travel patterns due to the new tolls. This could result in less frequent journeys, shifting modes, or in the most extreme cases, not travelling. After 2018, traffic flow behavior at Purgatorio toll plaza is relatively stable over the whole forecast, showing AADT annual growths of 2-4%. The above is illustrated by the following figure, which shows the population centers near Purgatorio toll plaza s new location. Results of traffic projections for this station are shown below. Figure 8.4: Difference in counted traffic in current and future locations of Purgatorio Source: Steer Davies Gleave Steer Davies Gleave Page 134 of 259

35 Traffic study for project Conexión Antioquia - Bolívar Figure 8.5: Base scenario AADT traffic projections Purgatorio Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 135 of 259

36 Traffic study for project Conexión Antioquia - Bolívar Table 8-3: Base scenario AADT traffic projections Purgatorio Note: Traffic projections of this station might vary according to the ultimate location of the toll plaza. These projections consider the location of Purgatorio according to what is stated in the concession contract for the complete calendar year of Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Please note that as of the date of this report there it is not certainty about the definite location of Purgatorio toll plaza, which will impact the total volume and the distribution among categories. In particular, the location Steer Davies Gleave Page 136 of 259

37 Traffic study for project Conexión Antioquia - Bolívar might determine bus and heavy vehicles volumes related to the location of quarries, construction materials depots, among other factors. Due to the fact that the future location of Purgatorio is closer to Montería, and considering what have happened recently with the installation of different toll plazas in Colombia, there is a possibility that the tariff structure of this station change with the implementation of special tariffs or discounts for frequent users. This might impact both, the traffic volume and the revenue projected on this plaza. It is important to place attention about the decisions that will be taken with respect to the location of Purgatorio and to review traffic projections once a final location for this station is defined. Los Cedros The results for this location are summarized as follows: In 2024/2025 there is an increase in total AADT of approximately 5%. This is due to the expected opening of an additional corridor linking the center and north of Antioquia with the Caribbean coast via Urabá. Part of this route includes the construction of the Toyo Tunnel which complements Mar1 and Mar2. This corridor represents a new alternative for long distance traffic between Medellín and the south and center of Antioquia, towards the zone of Montería. These trips will use the Los Cedros toll. Generally, however the traffic behavior is stable at this toll plaza, with AADT growth between 2% and 4% annually. Figure 8.6: Base scenario AADT traffic projections Los Cedros Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 137 of 259

38 Traffic study for project Conexión Antioquia - Bolívar Steer Davies Gleave Page 138 of 259 Table 8-4: Base scenario AADT traffic projections Los Cedros Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas.

39 Traffic study for project Conexión Antioquia - Bolívar San Carlos The results for this station are as follows: San Carlos has the lowest volumes of forecast traffics due it primarily serving long distance traffic, more than 60%, is forecast from This toll starts to operate in 2018 with the opening of this section of road. The complete San Carlos relief road of (El 15 - San Carlos - Cereté) is completed until Thus, an increase approximately 150 vehicles (150%) is expected from 2018 to The years 2020 and 2021 show an 85% increase in traffic (289 to 536 vehicles), due to the completion of IFUs 2, 6, and 7. These interventions improve the conditions of the regional road network between Planeta Rica, Montería, Cereté, Lorica, Coveñas and Tolú. In 2025 there is a 76% increase in total vehicle traffic due to the opening of the Córdoba - Sucre concession. This results in a diversion of trips from the competing routes. After 2025 all interventions in the road network in the Study Area are completed, so the growth stabilizes to between 2% and 3% per year due to macroeconomic behavior. Figure 8.7: Base scenario AADT traffic projections San Carlos Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 139 of 259

40 Traffic study for project Conexión Antioquia - Bolívar Steer Davies Gleave Page 140 of 259 Table 8-5: Base scenario AADT traffic projections San Carlos Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas.

41 Traffic study for project Conexión Antioquia - Bolívar Mata de Caña The results for this location are summarized below: Across all years, light vehicles have the largest share, up to 70% of total AADT. From 2022 to 2025 category VII freight vehicles rise from 2% to 8% of total AADT due to the opening of the competing Córdoba - Sucre concession and the increase in tolls on that route (see section 6.2). Therefore, there are significant increases in traffic in 2021 and 2025: In 2020/21 there is an increase in AADT of the order of 630 vehicles (21%), largely made up of an increase in category V to VII vehicles. A similar situation occurs in 2025, where the AADT increase is about 20% of total due to a growth of approximately 80% in V, VI and VII category vehicles. Figure 8.8: Base scenario AADT traffic projections Mata de Caña Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 141 of 259

42 Traffic study for project Conexión Antioquia - Bolívar Steer Davies Gleave Page 142 of 259 Table 8-6: Base scenario AADT traffic projections Mata de Caña Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas.

43 Traffic study for project Conexión Antioquia - Bolívar Caimanera The results for this location are summarized below: Significant traffic growth is observed after the completion of the Ruta al Mar network in 2021, of the order of 30%. The improvements in IFUs 2, 6, and 7 provide better conditions between Montería and Tolú. The composition of traffic changes in 2021 due to the increase in heavy vehicles. The heavy vehicle AADT share increases from approximately 4% to approximately 15% after Part of this forecast behavior is due to the increase in tariffs on competing routes of the Córdoba - Sucre concession. Light vehicles are the predominant category for this location, accounting for more than half of the total AADT. This is due to the strong local traffic component, between Tolú and Coveñas. Figure 8.9: Base scenario AADT traffic projections Caimanera Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 143 of 259

44 Traffic study for project Conexión Antioquia - Bolívar Steer Davies Gleave Page 144 of 259 Table 8-7: Base scenario AADT traffic projections Caimanera Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas.

45 Traffic study for project Conexión Antioquia - Bolívar San Onofre The behavior forecast at this toll plaza is very similar the two toll plazas previously analyzed. A 28% traffic growth (approximately 600 vehicles), is forecast when the Ruta al Mar network is completed in The introduction of the Córdoba - Sucre concession, including tariff increase, is forecast to divert traffic onto this route, increasing traffic by 9%, particularly for categories IV to VII. Figure 8.10: Base scenario AADT traffic projections San Onofre Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 145 of 259

46 Traffic study for project Conexión Antioquia - Bolívar Steer Davies Gleave Page 146 of 259 Table 8-8: Base scenario AADT traffic projections San Onofre Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas.

47 Traffic study for project Conexión Antioquia - Bolívar 9. SENSITIVITY ANALYSIS The baseline traffic projections presented in chapter 7 represent the best estimate by Steer Davies Gleave based on available information. However, there are certain assumptions that may change in the future and can be expected to have a direct influence on the traffic forecasts for the Ruta al Mar Concession. Consequently, a sensitivity analysis has been undertaken to test the impact of changing the key modeling assumptions. The sensitivities tested relate to: o o o changes in Colombia s GDP projections; long-term annual traffic growth; and variations to the external projects described in section of this report. Five different alternative sets of assumptions were considered and tested. These sensitivities were then combined with an analysis of the effects of each compared with the baseline scenario. The results of these sensitivities show a marked difference between regional and long distance traffic. In the first case, variations on the nation s economy mainly affects the toll plazas of the project, which are located on the northern and southern edges of the road, specifically between Caucasia and Planeta Rica, Montería and Arboletes, and Tolú and Cruz del Viso. The main traffic there is long distance between ports on the Caribbean and the interior of the country. On the other hand, regional and local traffic is more sensitive to the change in trip costs caused by variations in the road network and tolls. This is evident in the projections of toll plazas located between Monteria and Tolú (Caimanera, Mata de Caña and San Carlos). Traffic at these toll plazas is mainly frequent short distance trips, due to their proximity to populated centers. SCENARIO I (SI): From the general strategic analysis of the project we observed that the opening of the following projects result in a diversion of trips onto Ruta al Mar, partly due to their increased costs. Consequently, we analyze the impact of their completion in each timeframe. Concesión Córdoba Sucre Autopista Conexión Norte Autopista al Río Magdalena This scenario tests the impact on forecasted traffic should the above projects not be developed and therefore conditions remain as they currently are. Results and comparisons with the base scenario The effects of SI are shown below: There is no noticeable impact on traffic in the southern section of the project: La Apartada, Manguitos, and Purgatorio. The impacted long distance traffic is forced along the southern sections of the project regardless of whether the Córdoba - Sucre, Magdalena 2 and Conexión Norte projects exist. This traffic does not have alternate routes connecting the north of Antioquia with the coastal zone of the Caribbean. The tolls of Cedros and Purgatorio primarily serve local traffic around Montería and see minimal impact, as the projects tested in this sensitivity test primarily impact long distance trips. The traffic projected at the tolls located north of Planeta Rica (Mata de Caña, San Carlos, Caimanera, and San Onofre), are negatively affected in this test. This is mainly due to the absence of the Córdoba - Sucre Concession which is forecast to increase tolls on competing routes. The competing routes remain cheaper than Ruta al Mar and so the forecast trip diversion does not occur. Steer Davies Gleave Page 147 of 259

48 Traffic study for project Conexión Antioquia - Bolívar SCENARIO II (SII): This scenario is a variation from SI including the construction and operation of a port in the municipality of Turbo in Antioquia; a new facility being envisioned and studied at the time of production of this report. The operation of this port is expected to cause a diversion of heavy vehicle trips currently travelling between the Coffee Hub, Antioquia, and the port of Cartagena. As such, the impact is forecasted to be a reduction in long distance heavy vehicle trips using the concession. The status of the different port initiatives in the Uraba zone is still uncertain to the date of this report. Several terminals have requested concession agreements and environmental licences. Some ports have been awarded the concessions but contracts have not been signed at the time of producing this report. Monitoring of this initiatives is recommended by interested parties; the realisation of this scenario does not only depend on the ports being operational but also on the construction, operation and implementation of the work plan and planned tariff schemes for toll roads of Mar 1, Mar 2 and Tunel del Toyo. Impact of Puerto Antioquia on the project Projects are currently underway to develop the Puerto Antioquia, in the southeast of the Gulf of Urabá near Turbo, Antioquia. The port is expected to be connected to the main production and consumption centers, serving the Coffee Hub, Antioquia, and parts of the center of the country. The port is assumed to serve dry and refrigerated containers, dry bulk, general cargo, and vehicles 14. The potential port is expected to feature: Annual capacity of 1,200,000 TEUS Annual capacity of seven million tons in annual cargo. Perishable goods warehouses, for refrigerated and dry products, consolidation and deconsolidation goods Considering its proximity to Mar 1 and Mar 2, and the Toyo Tunnel; this section contains the analysis of the impact that such a port would have on heavy vehicle trips on the project roads. The methodology was as follows: i. Identification of main cargo sources using the Antioquia - Bolívar corridor destined for Cartagena port. ii. Comparison of generalized travel costs for cargo trucks (categories V, VI and VII) between the Cartagena port and Turbo port. iii. Analysis of the annual total cargo tonnage transported on the routes studied (base year ). iv. Analysis of foreign trade cargo using Cartagena port. v. Analysis of the total annual trips carried out on the routes studied (base year 2013). The long-distance trips between Medellín - Cartagena and Manizales - Cartagena that may migrate to Turbo were determined from the heavy vehicles survey (see section 3). These are the long-distance trips that may migrate to Turbo if the port facility is open, as they carry the largest amount of cargo among the project s routes. This is due to the reduction in the travel cost when reaching Turbo compared to Cartagena. The network model was used to analyses the generalized costs of large trucks on these routes in the 2025 base scenario (2025 was chosen as it is the year in which the relevant road network is complete and is when the alternate route to Urabá is consolidated with projects Mar1, Mar 2 and the Toyo Tunnel). The results are listed in the following Table. 14 Information extracted from 15 Information from Ministry s of Transport report, containing freight origin destination matrix of Colombia was considered in this study. Steer Davies Gleave Page 148 of 259

49 Traffic study for project Conexión Antioquia - Bolívar Table 9-1: Comparison of generalized travel costs for large trucks on the main routes toward port areas in Cartagena and Turbo (Base scenario, 2025) Route Medellín Cartagena $ 216,740 Medellín Puerto Turbo $ 86,500 Manizales Cartagena $ 275,352 Manizales Puerto Turbo $ 145,431 Source: Steer Davies Gleave Generalized travel cost (GTC) in the 2025 scenario (weekday) (2015 COP) Variation in GTC between destination turbo and Cartagena -60% -47% Figure 9.1: Generalized travel cost for origin destination trips of interest Source: Steer Davies Gleave Steer Davies Gleave Page 149 of 259

50 Traffic study for project Conexión Antioquia - Bolívar The previous table shows a marked reduction in the generalized cost of travel to Turbo compared to Cartagena is observed (60% and 47% for the OD pairs originated in Medellín and Manizales). This gives some evidence that there would be landside demand to use Turbo as a distribution port rather than Cartagena. The cargo tonnage in the Study Area is analyzed using the most recent information available from the Ministry of Transport. This information, completed in 2013, updates the cargo Origin-Destination matrix from which: Table 9-2: Total tons mobilized between origin and destination Origin Destination Total annual tons Cartagena Manizales 16,694 Cartagena Medellín 173,321 Manizales Cartagena 19,575 Medellín Cartagena 100,578 Total 310,169 Source: Steer Davies Gleave, based on the OD matric from the Ministry of Transportation (2013) The main products exported and imported through Cartagena Regional Port Company are listed below 16. Table 9-3: Main imported and exported product by the SPR Cartagena Main exported goods Main imported goods Several unspecified Several unspecified Other raw materials Other Chemicals Other Chemicals Paper and paperboard Ferronickel Other food products Construction materials General spare parts Other food products Polyethylene and plastic materials Bottles and containers Electrical apparatus and devices Paper and paperboard Construction materials Cork and wood manufactures (except furniture) Meat and meat preparations Raw hides and skins Tiles and ceramics Coffee excelso Vehicles & Automotive Smoke black Iron ore Polyethylene and plastic materials Lubricant oils Source: Steer Davies Gleave, based on information from the Superintendency of Ports and Transport Using this information, we analyzed the relevant product tonnage transported on the Medellín - Cartagena and Manizales - Cartagena OD pairs. Coal, oil and its derivatives, and chemical products were assumed to continue to use Cartagena Port based on the specialized industry for the treatment of these type of products there (particularly the recent opening of 16 Steer Davies Gleave Page 150 of 259

51 Traffic study for project Conexión Antioquia - Bolívar an oil refinery). It was assumed that 90% of relevant cargo is for foreign trade with the remaining share for local distribution. The following table lists the total annual tonnage for export transported between the OD pairs. Table 9-4: Total tonnage transported between Origin-Destination, export market product Origin Destination Total annual tons 90% of the total Cartagena Medellín 99,849 89,864 Manizales Cartagena Medellín Cartagena 24,902 22,412 Total 125, ,661 Source: calculated by Steer Davies Gleave, based on the OD matrix information from the Ministry of Transportation (2013) It is notable that the total is considerably lower than the future annual capacity of Turbo port (7 million tons per annum). The annual cargo movements were then converted into potential heavy vehicle trips which would no longer use Cartagena. The number of such trips is shown in the table below. Table 9-5: Total annual trips in large trucks Origin Destination Total annual trips Cartagena Manizales 549 Cartagena Medellín 5,955 Manizales Cartagena 723 Medellín Cartagena 5,933 Total 13,160 Source: calculated by Steer Davies Gleave, based on the OD matrix information from the Ministry of Transportation (2013) The annual trips above were then factored down to only represent the export market. Table 9-6: Total annual trips in large trucks with exporting cargo Origin Destination Total 90% of Total Cartagena Medellín 3,989 3,590 Manizales Cartagena Medellín Cartagena 1,317 1,185 Total 5,327 4,795 Source: elaborated by Steer Davies Gleave, based on the OD matrix information from the Ministry of Transportation (2013) From these results, it was estimated that 42.5% of the trips would no longer use Ruta al Mar if the Turbo port is constructed. This proportion is obtained as the relationship between travel Between Manizales, Medellín and the Port of Cartagena for selected foreign trade products and total travel between these pairs (4,795 trips / 1,12661 trips) Steer Davies Gleave Page 151 of 259

52 Traffic study for project Conexión Antioquia - Bolívar In summary, in SII Mar 1, Mar 2, the Toyo Tunnel, and a future port facility in Turbo are included, and thus, 42.5% heavy vehicle trips (categories V, VI and VII) destined for the Cartagena port would use the Turbo Port and no longer use Ruta al Mar. Results and comparison with base case The key results of the sensitivity tests are summarized below. There is a reduction in the forecast traffic across all toll plazas from This is the year when both Turbo Port is expected to be in operation and the network upgrades will be in place. The impact on traffic is greater at those locations which have a higher share of long distance trips. This is broken down by location as follows. o After 2025 San Carlos and Caimanera see lower forecasts (54% and 24% respectively) compared to the base case. Though the percentage change is large, these locations have a small number of trips and so the overall magnitude of the change is relatively small. o San Carlos, Mata de Caña, and San Onofre are locations with significant long distance large trucks traffic. These locations see the greatest change (averages of 14%, 16% and 12% after 2025). o The locations where only minor impacts are observed are the Los Manguitos, Purgatorio, and Cedros toll plazas where the differences from 2025 are less than 7%. This is due to the high proportion of local traffic at these locations. SCENARIO III (SIII): This sensitivity test assumes the new IP Córdoba - Sucre does not come to fruition and that the roads subject to improvements (via that project) remain as they are today. This applies not only to road infrastructure but to toll fare structure and location. This sensitivity scenario aims to study the impact of not considering the operation of the main competing concession (new IP Córdoba - Sucre) in the forecasts of the toll plazas within the project, keeping the current competing route s toll fares after Results and comparison with base case The main effects of SIII are summarized below: The forecast total traffic is unchanged between 2018 and 2024, but from 2025 (when Concession Córdoba - Sucre starts operating in the base scenario) traffic reduces by 4%. Light vehicles and buses differ by only of 2% against base case, as this scenario largely affects long distance traffic because Concession Córdoba - Sucre significantly rise toll fares for large trucks (Categories V, VI and VII). This causes more trucks to use the competing road rather than the project. From 2025 onwards, large truck traffic is reduced by 18% in average, due to the absence of causing the project to be more expensive than the competing road. SCENARIO IV (SIV): This scenario considers more optimistic economic conditions, adding 0.5% to the GDP base projections up to It also assumes 2.5% of annual traffic growth from 2036 to Results and comparison with base case At all the toll plazas of the Project the difference tends to increase as the time goes by due to the cumulative effect of GDP over the time and the 2.5% growth after Over the first years of the project s construction and operation ( ), projections of between 1 and 3% are seen. In following years, the difference increases, reaching over 22% in Steer Davies Gleave Page 152 of 259

53 Traffic study for project Conexión Antioquia - Bolívar SCENARIO V (SV): This scenario considers more pessimistic economic conditions, subtracting 0.5% of the GBP base projections up to 2035 and assuming 1.5% of annual traffic growth from years Results and comparison with base case The differences increase with time due to the cumulative effect of the GDP change. The assumption of 1.5% annual traffic growth for years reinforces the reductions. In the first year s variations are from -1 to -3%. Over the following years ( ), variations increase up to 10% lower. In the final years of the projection ( ), variations rise from 12 to 18% less. SUMMARY The traffic forecasts for the base case and each sensitivity test for the project toll plazas are graphed below. The percentage change between the base case and each sensitivity test is also shown. Results of these projections can be found in Annex I. The main results show that the projected traffic for the peripheral plazas; such as Caucasia - Monteria (La Apartada, Los Manguitos and Purgatorio), Monteria - Arboletes (Cedros) and Tolú - Cruz del Viso (San Onofre); react similarly to variations in the assumptions across the different sensitivity tests. The macroeconomic conditions of the country are the attributes that impact the traffic projections most at these plazas. The estimates show that the highest economic growth (SIV) scenario generates the upper band in the projections while the lowest GDP (SV) scenario results in the lowest. This is because the northern and southern edges of the project see the most long-distance traffic, which is that most directly affected by external trade and the national economy. The La Apartada and Los Manguitos toll plazas located in the south of the project area between Caucasia and Planeta Rica, have a significant position because they face a minor risk of new competition from other infrastructure projects other than the construction of the Turbo port. It is also expected that traffic demand at these plazas will not suffer major fluctuations over time because they are located on an existing segment of the road that has had reasonably stable demand in recent times. These stations, together with Purgatorio have the most of the concession transactions and see the biggest large truck demand. On the other hand, San Carlos, Mata de Caña and Caimanera toll plazas, located on the northern area of the Córdoba Department between Monteria and Tolú, capture mostly regional and local trips that are usually shorter and more frequent. Thus, the traffic projections at these plazas are more sensitive to changes in the road network in the study area. This effect is shown in the graphs where scenarios SI, SII and SIII end up with the lowest traffic estimates at these plazas, lower than scenario SV that assumes the worst situation for the country s economy. An important assumption that significantly affects the traffic on five of the eight toll plazas is the potential start of operation of a new port in Turbo, Antioquia. This is illustrated in the graphic that sums up the traffic volume of the entire concession. As it was analyzed in the results of scenario SII, the trip costs between the inland areas (Coffee Hub, Antioquia) and Turbo are lower than the ones required to connect those origins/destinations with the port in Cartagena (being 60% lower for Medellín and 47% lower for Manizales). That future scenario reduces the traffic forecasts for La Apartada, Manguitos, San Carlos, Mata de Caña, Caimanera and San Onofre to half of the value that they presented on the case with the lowest economic projections. Nevertheless, the status of the different port initiatives in the Uraba zone is still uncertain to the date of this report 17, and the assumptions considered in scenario SII depends on other assumptions that are not necessarily related to the 17 Several terminals have requested concession agreements and environmental licences. Some ports have been awarded the concessions but contracts have not been signed at the time of producing this report. Steer Davies Gleave Page 153 of 259

54 Traffic study for project Conexión Antioquia - Bolívar project, such as road interventions between Medellín and Urabá (Mar 1, Mar 2 and Tunel del Toyo), or the decision of shipping companies to move their operations and logistic centers to the new port infrastructure. We think is important to include this scenario mainly for the potential impact that it can have on the traffic in the project even though at the time of production of this study it is not clear when or how the port is going to be developed. Monitoring of this port is recommended by interested parties. Figure 9.2: Total traffic forecasts for all of the project toll plazas, for the base case and each sensitivity test (AADT) Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Table 9-7: Total traffic of the project and percentage change between the base case and each sensitivity test Year Base Case 24,251 25,137 26,060 28,978 29,973 31,102 Percentage change between sensitivities and base case SI SII SIII SIV SV 0% 0% 0% 2% -1% 0% 0% 0% 2% -2% 0% 0% 0% 3% -2% -1% -1% 0% 3% -2% -1% -1% 0% 4% -3% -1% -1% 0% 4% -3% Steer Davies Gleave Page 154 of 259

55 Traffic study for project Conexión Antioquia - Bolívar Year Base Case 32,186 35,105 36,207 37,344 38,517 39,728 40,978 42,232 43,526 44,860 46,235 47,654 48,602 49,569 50,556 51,562 52,588 53,635 54,703 55,792 56,903 58,036 59,192 60,370 61,573 62,800 Percentage change between sensitivities and base case SI SII SIII SIV SV -1% -1% 0% 5% -4% -6% -11% -4% 5% -4% -6% -11% -4% 6% -5% -6% -11% -4% 6% -5% -6% -11% -4% 7% -6% -5% -11% -4% 7% -6% -5% -11% -4% 8% -6% -5% -11% -4% 8% -7% -5% -11% -4% 9% -7% -5% -11% -4% 9% -8% -5% -11% -4% 10% -8% -5% -11% -4% 10% -9% -5% -11% -4% 11% -9% -5% -11% -4% 11% -9% -5% -11% -4% 12% -10% -5% -11% -4% 12% -10% -5% -11% -4% 13% -11% -5% -11% -4% 14% -11% -5% -11% -4% 14% -12% -5% -11% -4% 15% -12% -5% -11% -4% 15% -13% -5% -11% -4% 16% -13% -5% -11% -4% 16% -13% -5% -11% -4% 17% -14% -5% -11% -4% 18% -14% -5% -11% -4% 18% -15% Steer Davies Gleave Page 155 of 259

56 Traffic study for project Conexión Antioquia - Bolívar Year Base Case 64,051 65,327 66,628 67,956 69,310 70,691 Percentage change between sensitivities and base case SI SII SIII SIV SV -5% -11% -4% 19% -15% -5% -11% -4% 19% -15% -5% -11% -4% 20% -16% -5% -11% -4% 20% -16% -5% -11% -4% 21% -17% -5% -11% -4% 22% -17% Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 156 of 259

57 Traffic study for project Conexión Antioquia - Bolívar Figure 9.3: Total traffic forecasts for each of the project toll plazas, for the base case and each sensitivity test (AADT) Toll La Apartada Toll Los Manguitos Toll Purgatorio Toll San Carlos Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 157 of 259

58 Traffic study for project Conexión Antioquia - Bolívar Toll Cedros Toll Mata de Caña Toll Caimanera Toll San Onofre Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 158 of 259

59 Traffic study for project Conexión Antioquia - Bolívar 10.Overall findings and conclusions of the traffic study This section summarizes the main findings of the traffic study and highlights the most important aspects that should be considered when evaluating the traffic forecasts for the project Antioquia - Bolívar. First it is important to highlight that the project is part of one of the main corridors of the country, linking the border with Ecuador in the south with the ports in the Caribbean Sea as part of the main western national road corridor. Some segments of the project are part of the main trunk routes for internal and external trade. Traffic has been stable or growing in the last twelve years where existing toll plazas records were evaluated. Although toll plazas exist in the region, it is also likely that some negative reaction to new plazas and increased tariffs will occur as it did after new plazas were installed in 2016 when certain existing tolls experienced significant fare increases. These factors are considered in our analysis including the expected recovery after a period of time. However, the concessionaire has identified toll evasion in several toll plazas that has not been considered explicitly in our forecasts. No specific sector related to a single commodity or economic activity dominates the traffic in the project network but rather a combination of suburban, long and short distance tourism trips and long-distance freight trips make up most of the demand. All these elements were incorporated into our analysis and forecasts. Fieldwork revealed the important difference between the traffic composition of toll plazas located near urban centers and the ones located in rural areas. For the urban areas, traffic is composed of a high portion of light vehicles and motorcycles. The trips captured by those stations tend to be more frequent but shorter, and are near Monteria, Lorica, Coveñas and Tolú. Toll plazas that share these characteristics are Purgatorio, Mata de Caña, Cedros and Caimanera. For the rural areas, the segments of the project that are intercity connections have a higher demand of heavy vehicles, capturing long-distance trips associated with the movement between the inlands of the country and the ports in the Caribbean, specially Cartagena and Barranquilla. These trips are less frequent and have higher operational costs in terms of tolls and fuel. Results found in opinion surveys show that locations near urban zones are the most affected by fare increases or the installation of new plazas. It is important to mention that this part of the country, (from northern Monteria to the municipality of Tolú) is highly populated by people of low incomes, who are more sensitive to changes on their trip costs. Between 2020 and 2025, conditions of roads within the project are expected to improve with a consequent increase in the capture of long-distance trips compared to the first years of projections. Over this span of time, the conditions over the competing route are not expected to change. After 2025 the new concession IP Córdoba Sucre is assume to start operations, having an impact on the traffic projections of the central area of the concession ( San Carlos, Mata de Caña, Caimanera). The assumption is that a new toll is also considered charging new toll fare resulting in attracted traffic to the toll plazas mentioned above. IP Córdoba - Sucre is currently registered as a private initiative in the National Planning Department database and is under feasibility studies. Notwithstanding this registration there is no official documentation that guarantees the development as assumed and therefore a sensitivity test was included to measure the impact of a different network configuration. It is recommended that all interested parties follow up on the development of this project. This assumption also considers that the Carimagua toll plaza (currently under INVIAS operation) returns to the charging condition for trucks corresponding to the INVIAS prices again, after Also, one of the key downsides included in the traffic projections is the potential operation of a new port facility in Turbo. If this occurs it can have an important impact on the traffic projections of the toll plazas of the concession, particularly those concentrating most of the project s traffic volume (La Apartada and Manguitos). This port facility, together with the completion of an improved corridor through Urabá (considering completion of Mar1, Mar2 and Túnel del Toyo projects) can considerable reduce travel costs of heavy traffic reaching the Caribbean from the inside zones of the country. Notwithstanding that there is no official documentation that guarantees the development of the port, as far as we know concession permits have been awarded, some progress has been achieved in the designs and studies but no final contracts to execute Steer Davies Gleave Page 159 of 259

60 Traffic study for project Conexión Antioquia - Bolívar the facilities have been sigend therefore a sensitivity test was included to measure this impact. It is recommended that all interested parties follow up on the development of this project. The plazas that provide the greater contribution in terms of traffic to the project are La Apartada and Manguitos. These stations have nearly 50% of the total traffic of the concession and are located on a segment that has no competing local routes. Results show that on these toll plazas, the most important changes come from variations in the country s economy, as the scenarios with low and high GDP illustrate, or if new port facilities in Urabá impact the market of the existing dominant ports in Cartagena and Barranquilla. Other groups of toll plazas composed mainly by regional and local demand are the ones near urban areas and the ones located in the northern part of the Córdoba Department, such as San Carlos, Caimanera and Mata de Caña. The local and regional traffic that represents the largest portion of the demand on these plazas is less sensitive to economic variations and more sensitive to changes on road infrastructure and toll fares. Steer Davies Gleave Page 160 of 259

61 Traffic study for project Conexión Antioquia - Bolívar 11.ANNEX ANNEX A: TYPICAL CONDITIONS FOUND IN THE DATA COLLECTION Corridor Description Paved National bidirectional highway with two lanes per way direction and a wide central median. Pavement in good state as well as horizontal and vertical signage. Traced on plain flat terrain. Montería - Cereté: Route 21, route code: Paved National Highway Concession Córdoba - Sucre. Two-way road with one lane per way direction, pavement in good state and shoulder verge on both sides. Plain Flat terrain with horizontal and vertical signage in good state. Montería - Arboletes: Route 74, route code: Paved National Highway Concession Transversal de Las Américas. Corridor Córdoba Sucre Concession Paved road in good state, two-way with one lane per direction way. Horizontal and vertical signage in good state, narrow verge and traced over uneven terrain. Planeta Rica - La Y : Route 25, route code: Sector Caucasia - Planeta Rica INVIAS Paved highway with one lane per direction way. The pavement is in good state and it has a shoulder verge on each side. The terrain is mostly plain flat, with mild undulations. The second road is currently under construction. La Y - Sahagún: Ruta 25, route code: 2514 Concession Córdoba - Sucre. Steer Davies Gleave Page 161 of 259

62 Traffic study for project Conexión Antioquia - Bolívar Paved bidirectional road with shoulder verges and pavement in good state. Plain Flat terrain with mild undulations. Signage in good state. Sahagún - Chinú: Route 25, route code: 2514 Concession Córdoba - Sucre. Divided highway with two lanes per direction way in the section Corozal - Sincelejo. Between Sincelejo and Chinú, the road is bidirectional with only one lane per way. This section was traced travelled on a flat plain terrain with mild undulations. Pavement and signage in good state. Chinú - Corozal: Route 25, route code: 2514 Concession Córdoba - Sucre. Simple two-way road with one lane per direction way, with pavement in good state and horizontal and vertical signage in good conditions. Shoulder Verge with variable width for almost the whole route. Flat Plain terrain with some undulations. Corozal - Puerta de Hierro - El Carmen de Bolívar: Route 25, route code: Sector Puerta de Hierro - Carreto, between Puerta de Hierro - Corozal. Paved National Highway. Single two-way road with one lane per direction way. Plain topography with some undulations. The shoulder verge appears in different sections with variable width. High traffic of trucks and light vehicles. El Carmen de Bolívar - Campoalegre: Route 25, route code: Sector Puerta de Hierro - Carreto. Road with one lane per direction way without shoulder verges in most of the sections. Paved national highway in good state and with signage in good conditions. Plain Flat terrain with the presence of some undulations. Campoalegre - Cruz del Viso: Route 25, route code: 25BL02. Sector Carreto - Cruz del Viso Steer Davies Gleave Page 162 of 259

63 Traffic study for project Conexión Antioquia - Bolívar Other corridors: Bidirectional single lane road with no verge and pavement in good state. Mostly located in flat plain terrain with some undulations. Vertical and horizontal signage in good conditions. Campoalegre-Calamar: Paved National Highway Puerta de Hierro Concession - Palmar de Varela and Carreto - Cruz del Viso Two-way highway with one lane per direction way. Two more lanes to complete the divided highway are under construction. The pavement is in good state as well as the signage. The terrain is mostly flat plain. Sincelejo - Tolúviejo: Route 25, code 25SC01. Paved National Highway Concession Córdoba - Sucre. Divided highway with two lanes per direction way, pavement and signage in good state. Plain Flat terrain. Ciénaga de Oro - Cereté: Route 74, code Paved National Highway. Concession Córdoba - Sucre. Paved highway characterized by its two-way road with one lane per direction way. It does not have shoulders verges. In regular state with some sections with badly maintained pavement. Plain Flat terrain. Horizontal signage not visible in some segments. Lorica- San Andrés de Sotavento: Route 78, code Sector Lorica - Chinú. Route owned by INVIAS. Bidirectional road with one lane per direction way. Pavement in regular state and in some sections, inexistent horizontal signage. Vertical signage in acceptable conditions. Road with no verge shoulder and in mostly flat plain terrain. San Andrés de Sotavento - Chinú: Route 78, code Sector Lorica - Chinú. Route owned by INVIAS. Steer Davies Gleave Page 163 of 259

64 Traffic study for project Conexión Antioquia - Bolívar Paved National highway with single lane per direction way and wide shoulders verges. Pavement and signage in good state. Los Pantanos - Santa Lucía: Route 74, code 74CR02. Sector Santa Lucía - Moñitos. Route owned by INVIAS. National Highway, unpaved and owned by INVIAS. It is in bad state, with almost no vertical signage along the route. It has a variable width but it is mostly narrow. Santa Lucía - Corregimiento Bajo Blanco: Route 74, code 74CR02. Sector Santa Lucía - Moñitos. Route owned by INVIAS. Unpaved National Highway owned by INVIAS. Road in regular and bad state, with partial vertical signage and non-existent horizontal signage. Corregimiento Bajo Blanco - Moñitos: Route 74, code 74CR02. Sector Santa Lucía - Moñitos. Road with a poorly maintained pavement. It is bidirectional with one lane per direction way with no vergeshoulder. Signage in good state. Moñitos - San Bernardo del Viento: Route 90, code Sector Moñitos - Lorica. Route owned by INVIAS. Paved two-way road with one lane per direction way, with no verges shoulders, pavement in good state and good vertical and horizontal signage. Mostly flat plain terrain. San Bernardo del Viento - Lorica: Route 90, code Sector Moñitos - Lorica. Route owned by INVIAS. Source: Steer Davies Gleave Steer Davies Gleave Page 164 of 259

65 Traffic study for project Conexión Antioquia - Bolívar ANNEX B: DETAILED RESULTS OF SPEED MEASUREMENT Table 11-1: Average trip speed at the main corridors of the study area Segment Segment description Average speed (km/h) 1 Montería - Los Pantanos 76 2 Montería - Patio Bonito Patio Bonito - Planeta Rica Planeta Rica Planeta Rica - La Apartada Planeta Rica - La Ye La Ye - Ciénaga de Oro 78 8 Ciénaga de Oro - Cereté Cereté - Lorica Lorica - San Andrés de Sotavento San Andrés de Sotavento - Chinú Chinú - Sahagún Sahagún - La Ye Cereté - Montería Montería Los Pantanos - Santa Lucía Santa Lucia - Corregimiento Bajo Blanco Corregimiento Bajo Blanco Corregimiento Bajo Blanco - Moñitos Moñitos - San Bernardo del Viento San Bernardo del Viento - Lorica Tolú - San Onofre Tolúviejo - Sincelejo Tolú - Tolúviejo Coveñas - Tolú Lorica - Coveñas Tolúviejo - San Onofre San Onofre - María La Baja María La Baja - Cruz del Viso Carreto - Cruz del Viso Carmen de Bolívar - Carreto Corozal - Carmen de Bolívar Chinú - Corozal 66.7 Steer Davies Gleave Page 165 of 259

66 Traffic study for project Conexión Antioquia - Bolívar Segment Segment description Average speed (km/h) 34 Calamar - Carreto Source: Steer Davies Gleave Steer Davies Gleave Page 166 of 259

67 Traffic study for project Conexión Antioquia - Bolívar ANNEX C: TRAFFIC COUNTS RESULTS Master stations traffic counts per vehicle type and per day in each station The master counting stations that show more traffic are Sahagún and Cruz del Viso. The same vehicles were identified passing various times by the toll at the station Sahagún, transporting different passengers informally between Sahagún, Chinú and Sincelejo, which explains the large traffic volumes registered at this station. Additionally, there is a high traffic level at the location of the future toll plaza at Purgatorio. This could be related to the urban traffic that is observed in the area nearby to Montería. San Antero and San Onofre stations were the locations that registered the smallest traffic volumes. This could be caused by the presence of competing routes parallel to these stations, which have better driving conditions. The table below shows the distribution by category and by day of the week for the counted vehicles at each station. Steer Davies Gleave Page 167 of 259

68 Traffic study for project Conexión Antioquia - Bolívar Table 11-2: Vehicle volumes per type of vehicle per day master stations ID Station Day Light vehicles Buses Small trucks Large trucks Motor cycles Bicycles Total* M1 La Apartada Monday 1, , ,408 M1 La Apartada Tuesday 1, ,065 1,077 2, ,941 M1 La Apartada Wednesday 1, ,154 1,237 2, ,193 M1 La Apartada Thursday 2, ,279 1,389 3, ,919 M1 La Apartada Friday 1, ,124 1,37 3, ,647 M1 La Apartada Saturday 1, , , ,361 M1 La Apartada Sunday 1, ,058 2, ,858 M1 La Apartada Total* 12,321 1,746 7,471 7,789 20, ,327 M1 La Apartada Daily average* 1, ,067 1,113 2, ,190 M2 Purgatorio Monday 1, , ,001 M2 Purgatorio Tuesday 1, , ,088 M2 Purgatorio Wednesday 1, , ,14 M2 Purgatorio Thursday 1, , ,226 M2 Purgatorio Friday 2, , ,45 M2 Purgatorio Saturday 1, , ,109 M2 Purgatorio Sunday 1, , ,248 M2 Purgatorio Total* 13,162 2,974 4,015 1,111 17, ,262 M2 Purgatorio Daily average* 1, , ,039 M3 Cedros Monday 1, , ,1 M3 Cedros Tuesday 1, , ,143 M3 Cedros Wednesday 1, , ,176 M3 Cedros Thursday 1, , ,345 M3 Cedros Friday 1, , ,325 M3 Cedros Saturday 1, , ,488 M3 Cedros Sunday 1, , ,835 Steer Davies Gleave Page 168 of 259

69 Traffic study for project Conexión Antioquia - Bolívar ID Station Day Light vehicles Buses Small trucks Large trucks Motor cycles Bicycles Total* M3 Cedros Total* 8,803 2,048 3, , ,412 M3 Cedros Daily average* 1, , ,202 M4 Carimagua Monday , ,898 M4 Carimagua Tuesday , ,236 M4 Carimagua Wednesday , ,457 M4 Carimagua Thursday , ,741 M4 Carimagua Friday ,076 1, ,658 M4 Carimagua Saturday , ,429 M4 Carimagua Sunday , ,239 M4 Carimagua Total* 4, ,461 6,639 7, ,658 M4 Carimagua Daily average* , ,380 M5 Mata de Caña Monday 1, , ,571 M5 Mata de Caña Tuesday 1, , ,794 M5 Mata de Caña Wednesday 1, , ,759 M5 Mata de Caña Thursday 1, , ,172 M5 Mata de Caña Friday 1, , ,033 M5 Mata de Caña Saturday 2, , ,236 M5 Mata de Caña Sunday 1, , ,531 M5 Mata de Caña Total* 12,682 3,006 3,26 1,148 15, ,096 M5 Mata de Caña Daily average* 1, , ,873 M6 La Esperanza Monday 1, , ,588 M6 La Esperanza Tuesday 1, , ,743 M6 La Esperanza Wednesday 1, , ,846 M6 La Esperanza Thursday 1, ,003 2, ,952 M6 La Esperanza Friday 1, ,075 3, ,293 M6 La Esperanza Saturday 2, , ,017 Steer Davies Gleave Page 169 of 259

70 Traffic study for project Conexión Antioquia - Bolívar ID Station Day Light vehicles Buses Small trucks Large trucks Motor cycles Bicycles Total* M6 La Esperanza Sunday 1, , ,904 M6 La Esperanza Total* 12,963 2,873 4,582 5,925 18, ,343 M6 La Esperanza Daily average* 1, , ,763 M7 San Onofre Monday ,806 M7 San Onofre Tuesday ,823 M7 San Onofre Wednesday ,058 M7 San Onofre Thursday ,246 M7 San Onofre Friday ,407 M7 San Onofre Saturday ,385 M7 San Onofre Sunday ,539 M7 San Onofre Total* 4,69 1,638 3,857 4,079 3, ,264 M7 San Onofre Daily average* ,038 M8 El Carmen Monday 1, , ,174 M8 El Carmen Tuesday 1, , , ,804 M8 El Carmen Wednesday 1, , , ,92 M8 El Carmen Thursday 1, , , ,871 M8 El Carmen Friday 1, , , ,254 M8 El Carmen Saturday 1, , ,519 M8 El Carmen Sunday 1, , ,745 M8 El Carmen Total* 10,995 2,508 7,383 5,401 25,244 1,313 26,287 M8 El Carmen Daily average* 1, , , ,755 M9 Sahagún Monday 2, , , ,345 M9 Sahagún Tuesday 2, ,136 1,153 3, ,571 M9 Sahagún Wednesday 2, ,323 1,46 3, ,249 M9 Sahagún Thursday 3, ,51 1,476 3, ,891 M9 Sahagún Friday 3, ,414 1,425 7, ,235 Steer Davies Gleave Page 170 of 259

71 Traffic study for project Conexión Antioquia - Bolívar ID Station Day Light vehicles Buses Small trucks Large trucks Motor cycles Bicycles Total* M9 Sahagún Saturday 3, , , ,795 M9 Sahagún Sunday 2, , ,68 M9 Sahagún Total* 21,95 3,181 8,291 8,344 30,112 1,298 41,766 M9 Sahagún Daily average* 3, ,184 1,192 4, ,064 M10 San Antero Monday , ,487 M10 San Antero Tuesday , ,48 M10 San Antero Wednesday , ,472 M10 San Antero Thursday , ,754 M10 San Antero Friday , ,802 M10 San Antero Saturday 1, , ,28 M10 San Antero Sunday 1, , ,907 M10 San Antero Total* 7, ,945 1,003 16, ,182 M10 San Antero Daily average* 1, , ,74 M11 Cruz Del Viso Monday 1, ,657 1, ,582 M11 Cruz Del Viso Tuesday 1, ,118 1, ,576 M11 Cruz Del Viso Wednesday 1, ,98 1, ,298 M11 Cruz Del Viso Thursday 1, ,205 1,955 1, ,49 M11 Cruz Del Viso Friday 1, ,156 1, ,688 M11 Cruz Del Viso Saturday 2, ,106 2,145 1, ,239 M11 Cruz Del Viso Sunday 1, ,283 1, ,081 M11 Cruz Del Viso Total* 12,92 4,292 6,625 13,117 8, ,954 M11 Cruz Del Viso Daily average* 1, ,874 1, ,279 M12 Campoalegre Monday 1, ,571 M12 Campoalegre Tuesday ,64 M12 Campoalegre Wednesday ,612 M12 Campoalegre Thursday ,838 Steer Davies Gleave Page 171 of 259

72 Traffic study for project Conexión Antioquia - Bolívar ID Station Day Light vehicles Buses Small trucks Large trucks Motor cycles Bicycles Total* M12 Campoalegre Friday 1, ,836 M12 Campoalegre Saturday 1, ,866 M12 Campoalegre Sunday ,966 M12 Campoalegre Total* 7,14 1,43 5,623 4,136 2, ,329 M12 Campoalegre Daily average* 1, ,618 M13 Manguitos C Monday 2, , ,913 M13 Manguitos C Tuesday 1, ,051 1,112 2, ,449 M13 Manguitos C Wednesday 2, ,391 1,146 2, ,001 M13 Manguitos C Thursday 2, ,343 1,347 2, ,298 M13 Manguitos C Friday 2, ,403 1,177 2, ,471 M13 Manguitos C Saturday 2, ,005 1,098 2, ,661 M13 Manguitos C Sunday 1, , ,892 M13 Manguitos C Total* 14,722 2,72 8,114 7,885 16, ,441 M13 Manguitos C Daily average* 2, ,107 1,071 2, ,669 M14 Purgatorio F Monday 2, , ,952 M14 Purgatorio F Tuesday 2, , ,071 M14 Purgatorio F Wednesday 3, , ,195 M14 Purgatorio F Thursday 3, , ,441 M14 Purgatorio F Friday 3, , ,752 M14 Purgatorio F Saturday 3, , ,425 M14 Purgatorio F Sunday 3, , ,336 M14 Purgatorio F Total* 22,5 5,443 5,556 2,888 27,531 1,223 36,387 M14 Purgatorio F Daily average* 3, , ,167 *The calculation of the total and average daily volumes does not include either motorcycles or bicycles Source: Steer Davies Gleave Steer Davies Gleave Page 172 of 259

73 Traffic study for project Conexión Antioquia - Bolívar Hourly profiles of master stations Hourly Profiles Station M1 La Apartada Weekday Saturday Sunday Hourly Profiles Station M2 Purgatorio Weekday Saturday Sunday Hourly Profiles Station M3 Cedros Weekday Saturday Sunday Steer Davies Gleave Page 173 of 259

74 Traffic study for project Conexión Antioquia - Bolívar Hourly Profiles Station M4 Carimagua Weekday Saturday Sunday Hourly Profiles Station M5 Mata de Caña Weekday Saturday Sunday Hourly Profiles Station M6 La Esperanza Weekday Saturday Sunday Steer Davies Gleave Page 174 of 259

75 Traffic study for project Conexión Antioquia - Bolívar Hourly Profiles Station M7 San Onofre Weekday Saturday Sunday Hourly Profiles Station M8 El Carmen Weekday Saturday Sunday Hourly Profiles Station M9 Sahagún Weekday Saturday Sunday Steer Davies Gleave Page 175 of 259

76 Traffic study for project Conexión Antioquia - Bolívar Hourly Profiles Station M10 San Antero Weekday Saturday Sunday Hourly Profiles Station M11 Cruz del Viso Weekday Saturday Sunday Hourly Profiles Station M12 Campoalegre Weekday Saturday Sunday Source: Steer Davies Gleave Steer Davies Gleave Page 176 of 259

77 Traffic study for project Conexión Antioquia - Bolívar Vehicular distribution at master stations M1 Toll plaza La Apartada M2 Toll plaza Purgatorio M3 Toll plaza Cedros M4 Toll plaza Carimagua M5 Toll plaza Mata de Caña M6 Toll plaza La Esperanza Steer Davies Gleave Page 177 of 259

78 Traffic study for project Conexión Antioquia - Bolívar M7 Toll plaza San Onofre M8 Toll plaza El Carmen M9 Sahagún M10 San Antero M11 Cruz del Viso M12 Campoalegre Steer Davies Gleave Page 178 of 259

79 Traffic study for project Conexión Antioquia - Bolívar M13 Manguitos C M14 Purgatorio F Source: Steer Davies Gleave Specific stations traffic counts per vehicle type and per day in each station ID Station Day Light vehicles Buses Small trucks Large trucks Motor cycles Bicycles Total* E1 Planeta Rica Tuesday 2, , ,884 E1 Planeta Rica Wednesday 2, ,136 1,016 2, ,23 E1 Planeta Rica Thursday 3, ,138 1,21 4, ,708 E1 Planeta Rica Saturday 2, , ,016 E1 Planeta Rica Sunday 2, , ,08 E1 Planeta Rica Daily average 2, , ,984 E2 Cereté/Ci énaga de Tuesday 5, , , ,024 Oro E2 Cereté/Ci énaga de Thursday 5, , ,866 Oro E2 Cereté/Ci énaga de Saturday 5, , ,466 Oro E2 Cereté/Ci énaga de Sunday 3, , ,641 Oro E2 Cereté/Ci énaga de Daily average* 4, , ,249 Oro E3 Lorica/Chi nú Tuesday 1, , ,169 E3 Lorica/Chi nú Thursday 1, , ,383 E3 Lorica/Chi Saturday 2, , ,425 Steer Davies Gleave Page 179 of 259

80 Traffic study for project Conexión Antioquia - Bolívar ID Station Day Light vehicles Buses Small trucks Large trucks Motor cycles Bicycles Total* nú E3 Lorica/Chi nú Sunday 1, , ,718 E3 Lorica/Chi nú Daily average* 1, , ,174 E4 Sincelejo Tuesday 5, ,509 1,134 7, ,453 E4 Sincelejo Thursday 6, ,71 1,367 7, ,032 E4 Sincelejo Saturday 5, ,427 1,098 7, ,429 E4 Sincelejo Sunday 4, , ,139 E4 Sincelejo Daily average* 5, ,356 1,096 6, ,013 E5 E5 E5 E5 Carmen de Bolívar Carmen de Bolívar Carmen de Bolívar Carmen de Bolívar Tuesday , ,061 Thursday ,07 1, ,299 Saturday ,093 1, ,324 Sunday ,008 1, ,968 E5 Carmen de Bolívar Daily average* ,028 1, ,163 E6 Tolú Tuesday 1, , ,962 E6 Tolú Thursday 1, , ,1 E6 Tolú Saturday 1, , ,242 E6 Tolú Sunday 1, , ,916 E6 Tolú Daily average* 1, , ,055 E7 Planta Argos Tuesday 1, , ,945 E7 Planta Argos Thursday 2, , ,109 E7 Planta Argos Saturday 3, , ,019 E7 Planta Argos Sunday 3, , ,23 E7 Planta Argos Daily average* 2, , ,576 E8 Traslado Purgatorio Tuesday 2, , ,436 E8 E8 Traslado Purgatorio Traslado Purgatorio Thursday 2, , ,509 Saturday 3, , ,629 Steer Davies Gleave Page 180 of 259

81 Traffic study for project Conexión Antioquia - Bolívar ID Station Day Light vehicles Buses Small trucks Large trucks Motor cycles Bicycles Total* E8 E8 Traslado Purgatorio Traslado Purgatorio Sunday 2, , ,512 Daily average* 2, , ,272 *The calculation of the total and average daily volumes does not include either motorcycles or bicycles **Average considering only the days were counts were undertaken Source: Steer Davies Gleave Vehicular distribution at specific stations E1 Planeta Rica E2 Cereté/Ciénaga de Oro E3 Lorica/Chinú E4 Sincelejo Steer Davies Gleave Page 181 of 259

82 Traffic study for project Conexión Antioquia - Bolívar E5 Carmen de Bolívar E6 Tolú E7 Planta Argos E8 Traslado Purgatorio Source: Steer Davies Gleave Steer Davies Gleave Page 182 of 259

83 Traffic study for project Conexión Antioquia - Bolívar Hourly volume profiles at specific stations Hourly Profiles Station E1 Planeta Rica Weekday Saturday Sunday Hourly Profiles Station E2 Cereté/Ciénaga de Oro Weekday Saturday Sunday Hourly Profiles Station E3 Lorica/Chinú Weekday Saturday Sunday Steer Davies Gleave Page 183 of 259

84 Traffic study for project Conexión Antioquia - Bolívar Hourly Profiles Station E4 Sincelejo Weekday Saturday Sunday Hourly Profiles Station E5 Carmen de Bolívar Weekday Saturday Sunday Hourly Profiles Station E6 Tolú Weekday Saturday Sunday Steer Davies Gleave Page 184 of 259

85 Traffic study for project Conexión Antioquia - Bolívar Hourly Profiles Station E7 Planta Argos Weekday Saturday Sunday Hourly Profiles Station E8 Traslado Purgatorio Weekday Saturday Sunday Source: Steer Davies Gleave Steer Davies Gleave Page 185 of 259

86 Traffic study for project Conexión Antioquia - Bolívar Count results of the toll avoidance routes Table 11-3: Vehicle volumes registered by the Concessionaire in the possible avoidance routes La Apartada Station Day Light vehicles Buses and small trucks Big trucks La Apartada Monday La Apartada Tuesday La Apartada Wednesday La Apartada Thursday La Apartada Friday La Apartada Saturday La Apartada Sunday La Apartada ADWT La Apartada Total Source: Steer Davies Gleave with data from the Concessionaire Table 11-4: Vehicle volumes registered by the Concessionaire in the possible avoidance routes Caimanera Station Day Light vehicles Buses and small Big trucks trucks Caimanera Monday Caimanera Tuesday Caimanera Wednesday Caimanera Thursday Caimanera Friday Caimanera Saturday Caimanera Sunday Caimanera ADWT Caimanera Total Source: Steer Davies Gleave with data from the Concessionaire Table 11-5: Vehicle volumes registered by the Concessionaire in the possible avoidance routes Los Manguitos Station Day Light vehicles Buses and small Big trucks trucks Manguitos Monday Manguitos Tuesday Manguitos Wednesday Manguitos Thursday Manguitos Friday Manguitos Saturday Manguitos Sunday Manguitos ADWT Manguitos Total Source: Steer Davies Gleave with data from the Concessionaire Steer Davies Gleave Page 186 of 259

87 Traffic study for project Conexión Antioquia - Bolívar ANNEX D: ORIGIN DESTINATION SURVEYS RESULTS Below is a summary of the surveys by location, direction, and day. All values presented below relate only to fully completed surveys. In other words, where it was possible to identify the origin, destination and the reason for travel (necessary to segment the demand) for light vehicles, and origin, destination and type of vehicle (number of axles) for trucks. It is noticeable in the tables below that there are slightly more surveys for heavy vehicles than for light vehicles. Regarding the days of the surveys, it is observed that for light vehicles there more surveys done for light vehicles during the weekends, particularly on Sundays. For cargo vehicles, the contrary occurred, the days that had the most surveys were the weekdays. This corresponds to the results in the vehicular count process described above, where there is a clear pattern of the volume being higher for light vehicles on the final days of the week and the weekends for both master stations and specific stations. For heavy vehicles, the volume registered is higher on weekdays then on weekends. Table 11-6: Origin-Destination survey per station Station Valid survey lights Valid surveys trucks La Apartada (EOD1) 1,301 2,107 Purgatorio (EOD2) 1,66 1,026 Cedros (EOD3) 1, Carimagua (EOD4) 495 1,467 Mata de Caña (EOD5) 1, La Esperanza (EOD6) 1, San Onofre (EOD7) 676 1,127 El Carmen (EOD8) 1,165 1,141 Sahagún (EOD9) 1,004 1,053 San Antero (EOD10) Cruz del Viso (EOD11) 920 1,275 Campoalegre (EOD12) Total 12,949 13,121 Source: Steer Davies Gleave Table 11-7: Origin-Destination survey per station and day Light vehicles Station Direction Thursday 19/11 Saturday 21/11 Sunday 22/11 La Apartada (EOD1) Montería - Medellín La Apartada (EOD1) Medellín - Montería Purgatorio (EOD2) Montería - Planeta Rica Purgatorio (EOD2) Planeta Rica - Montería Cedros (EOD3) Arboletes - Montería Cedros (EOD3) Montería - Arboletes Carimagua (EOD4) Sahagún - Planeta Rica Tuesday 24/11 Steer Davies Gleave Page 187 of 259

88 Traffic study for project Conexión Antioquia - Bolívar Station Direction Thursday 19/11 Saturday 21/11 Sunday 22/11 Tuesday 24/11 Carimagua (EOD4) Planeta Rica - Sahagún Mata de Caña (EOD5) Lorica - Cerete Mata de Caña (EOD5) Cerete - Lorica La Esperanza (EOD6) Tolú - Sincelejo La Esperanza (EOD6) Sincelejo - Tolú San Onofre (EOD7) Cartagena - Montería San Onofre (EOD7) Montería - Cartagena El Carmen (EOD8) Carmen De Bolívar - Sincelejo El Carmen (EOD8) Sincelejo - Carmen De Bolívar Sahagún (EOD9) Sincelejo - Sahagún Sahagún (EOD9) Sahagún - Sincelejo San Antero (EOD10) Coveñas - Lorica San Antero (EOD10) Lorica - Coveñas Cruz del Viso (EOD11) Cartagena - Malagana Cruz del Viso (EOD11) Malagana - Cartagena Campoalegre (EOD12) Barranquilla - Carreto Campoalegre (EOD12) Carreto - Barranquilla Total 2,828 3,491 3,594 3,036 Source: Steer Davies Gleave Table 11-8: Origin-Destination survey per station and day Trucks Station Direction Thursday 19/11 Saturday 21/11 Sunday 22/11 Tuesday 24/11 La Apartada (EOD1) Montería - Medellín La Apartada (EOD1) Medellín - Montería Purgatorio (EOD2) Montería - Planeta Rica Purgatorio (EOD2) Planeta Rica - Montería Cedros (EOD3) Arboletes - Montería Cedros (EOD3) Montería - Arboletes Carimagua (EOD4) Sahagún - Planeta Rica Carimagua (EOD4) Planeta Rica - Sahagún Mata de Caña (EOD5) Lorica - Cerete Mata de Caña (EOD5) Cerete - Lorica La Esperanza (EOD6) Tolú - Sincelejo La Esperanza (EOD6) Sincelejo - Tolú San Onofre (EOD7) Cartagena - Montería Steer Davies Gleave Page 188 of 259

89 Traffic study for project Conexión Antioquia - Bolívar Station Direction Thursday 19/11 Saturday 21/11 Sunday 22/11 San Onofre (EOD7) Montería - Cartagena El Carmen (EOD8) Carmen De Bolívar - Sincelejo El Carmen (EOD8) Sincelejo - Carmen De Bolívar Sahagún (EOD9) Sincelejo - Sahagún Sahagún (EOD9) Sahagún - Sincelejo San Antero (EOD10) Coveñas - Lorica San Antero (EOD10) Lorica - Coveñas Cruz del Viso (EOD11) Cartagena - Malagana Cruz del Viso (EOD11) Malagana - Cartagena Campoalegre (EOD12) Barranquilla - Carreto Campoalegre (EOD12) Carreto - Barranquilla Tuesday 24/11 Total 3,570 3,392 2,655 3,504 Source: Steer Davies Gleave Survey results The following subchapters summarize some of the results obtained in the surveys. For each type of vehicle (automobile and truck) there are indicators to: Reason for travel Frequency of trip Trip duration Cost Type of cargo The information is divided for weekday and weekend, in order to understand the different patterns in the periods of study. It is important to understand that the periods studied show profiles and segments of demand that vary in a significant way during the days analyzed. Light vehicles Reason for travel To know the behavior of the vehicular flow on the road, it is important to identify the main reasons behind users decisions to travel on a given route. Due to this, the proportion of reasons for travel per station, per type of day, were calculated and are shown in the following graphs: Steer Davies Gleave Page 189 of 259

90 Traffic study for project Conexión Antioquia - Bolívar Figure 11.1: Proportion of trip purpose in each station Light vehicles Week Day Weekend Source: Steer Davies Gleave The reasons for travel were grouped in such a way that the Hogar (Home) category includes all trips with the purpose of home and study, the category of Trabajo (Work) includes all trips with the purpose of work and places related to work or business, and Otros (other) category includes all trips that access services such as health, shopping and entertainment. For weekday, it is clear that the trips with the motive of work make up approximately half of the partition, and the trips that are related to home make up 30% of the total. In the station Sahagún (M9), there is a different behavior, most of the trips have a motive for home, followed by other with 30% of the partition. As expected, on weekend days the amount of trips done for work and business are reduced significantly, going down to 30%, while the trips done for tourism and others go up to an average of 40%. Travel frequency The frequency of their trip was asked. This is used to estimate the proportion of frequent users and occasional users. The following figure shows the percentage of participation for each frequency collected by type of day. Steer Davies Gleave Page 190 of 259

91 Traffic study for project Conexión Antioquia - Bolívar Figure 11.2: Proportion of trip frequency in each station Light vehicles Week Day Weekend Source: Steer Davies Gleave Travel frequency is variable between stations. The stations M5 (Mata de Caña), M6 (La Esperanza), and M9 (Sahagún), have a high proportion of regular users. This is due to the proximity of these stations to important urban centers in the area: Cereté - Lorica for station M5, Lorica for M6, and Sahagún for M9. For the latter regular users represent over 50% of the trips. As expected, the stations located further from urban centers registered less frequent trips, as is the case of San Onofre (M7) and the toll plaza Carimagua (M4), where these trips represent 10% of the total. Occasional trips are the majority in La Apartada, while weekly is the majority for Cruz del Viso. On weekends, the proportion of occasional trips raises almost by 10% while daily traffic falls. Travel time The trip duration is a way to measure if the project roads are used for short, medium, or long distance trips. The following figure shows the distribution of the trips according to their duration for weekdays and weekends. Figure 11.3: Proportion of travel times in each station Light vehicles Week Day Weekend Source: Steer Davies Gleave For this indicator, there were no significant differences between weekday and weekend. However, there were large differences between stations. Short trips, less than two hours, are the majority at La Apartada, Cedros and la Esperanza tolls, and the stations of Sahagún and San Antero. For the toll plazas that are further from urban Steer Davies Gleave Page 191 of 259

92 Traffic study for project Conexión Antioquia - Bolívar areas, such as San Onofre and El Carmen, and the count station Campoalegre, the predominant travel duration is between 2 and 4 hours. Travel Cost The survey asked about travel costs related to each vehicle. The following figure shows the average of this cost, separated by toll costs and fuel costs, per type of day and station. Figure 11.4: Average toll and gas costs in each station Light vehicles Week Day Weekend Source: Steer Davies Gleave The stations that registered higher generalized costs, for tolls and fuel, were Campoalegre and Carimagua. The station that showed the lowest costs was La Esperanza. Regarding the person responsible for those costs, the figures show that across all stations the driver of the vehicle is responsible for the costs. Trucks Type of cargo The type of cargo is an indicator of the possible sources of the attraction and generations of trips in the Study Area. The figures below show the distribution of the different types of cargo among the trucks surveyed. Steer Davies Gleave Page 192 of 259

93 Traffic study for project Conexión Antioquia - Bolívar Figure 11.5: Proportion of cargo type in each station Week Day Weekend Source: Steer Davies Gleave In general, there were not many differences observed in the data when compares weekdays and weekends. On average, a fourth of the products were perishable, such as agricultural and livestock. Construction materials also represent an important percentage of products, between 15% and 20% in average. Travel frequency The following figures show the travel distributions of cargo relative to trip frequency. Figure 11.6: Proportion travel frequency in each station Trucks Week Day Weekend Source: Steer Davies Gleave The following points regarding the frequency of truck trips can be concluded: Truck trips are repetitive. As shown in the figure, most of the trips have a daily or weekly frequency. La Esperanza and Sahagún have the highest share of daily trips; while the Carimagua station register most of its trips as weekly. This represents the short distance trips, related to urban centers, and long distance trips for locations that are further form the urban centers. On the other hand, Purgatorio has mostly weekly and occasional trips. Steer Davies Gleave Page 193 of 259

94 Traffic study for project Conexión Antioquia - Bolívar Travel time Trips by trucks, have longer travel times that those by light vehicles, since they generally have longer distances. The following tables show the distribution for these durations. Figure 11.7: Proportion of travel times in each station Trucks Week Day Weekend Source: Steer Davies Gleave La Apartada, Carimagua and Cruz del Viso, show that more than half of the surveyed trips are longer than 12 hours. This clearly shows that the long-distance trips use stations that are far from urban centers. For Cruz del Viso, long distance trips recorded are associated to its proximity to Cartagena port. Short distance trips, less than two hours, are predominant at the toll plazas of Mata de Caña and La Esperanza. Sahagún Station has mostly 2 and 8 hour trips. The stations that are further from main urban centers have more long distance trips. The contrary is true for toll plazas like Purgatorio, Cedros, La Esperanza, Mata de caña and the station San Antero, which have more short distance trips, since they are located closer to urban areas. Travel Cost The following figure shows an average of travel costs, separated by toll costs and fuel costs, by type of day and by station. It is important to keep in mind that these costs are estimated by the respondents at the moment of the survey. Steer Davies Gleave Page 194 of 259

95 Traffic study for project Conexión Antioquia - Bolívar Figure 11.8: Average toll and gas costs in each station Trucks Week Day Weekend Source: Steer Davies Gleave On average, the cost of fuel represents 78% of the total cost of the trip. The highest costs were registered at Carimagua and Cruz del Viso, which is consistent with the travel times recorded. Origin-Destination survey summary Light vehicles For light vehicles, on average, the trips with a work purpose are approximately half of the participation and that trips related to home compose about 30% of all weekday trips. As expected, during weekends the portion of work related trips drops to an average of 30%, while the tourism and other trips are 40% of the total. Regarding the frequency of the trips, for stations that are near urbanized areas most of the responses are daily and more frequent trips. The stations that are in areas farther from urban zones, record more occasional trips. The more frequent trips are carried out during the week, due to their work-related motive. The stations that are further from urban centers, record higher operational costs. This is explained by the fact that these stations capture more long distance trips. Trucks Most of the trips are daily or weekly. The stations with the most daily trips are La Esperanza and Sahagún; while the station that recorded most of the weekly trips was Carimagua. Stations that are farther from urban centers captured longer trips and therefore are more costly trips, such is the case of Cruz del Viso Steer Davies Gleave Page 195 of 259

96 Traffic study for project Conexión Antioquia - Bolívar ANNEX E: RESULTS OF THE STATED PREFERENCES SURVEYS Stated preference survey design Questionnaire The questionnaire was formed by five groups of questions: Filter: Question intended to identify the respondent s profile i. Questions about the trip made: Origin and destination of final trip Type of vehicle: lightweight or truck Characteristics of trip: purpose, who is responsible for associated costs, frequency, total travel time. Additionally, in the case of trucks, the owner and responsibility for deciding the route is investigated. ii. A stated preference exercise is performed for route selection Presentation slides of the project and explanation of stated preference exercise. Stated preference exercise, consisting of 9 cards (route used versus project). iii. Questions about the use of electronic toll system iv. Question to ask for user characteristics Gender Age Monthly income in the home Social-economic level v. Free comments Design of the stated preference exercise A generic stated preference survey for route selection was designed, between the current route and the potential future tolled route. The design was differentiated by the type of vehicle. The respondents had 9 questions, each of them with two alternatives to consider. Alternative 1 Current route Alternative 2 A new road Each alternative was characterized by travel time and costs. The differences between time and costs for each exercise have an implicit value of time. The answers recorded from all the exercises are a key input into a discreet choice model (Multinomial Logit Model) that quantifies value of time. The values corresponding to Cost Delta and Time Delta used in the SP survey design, and that are shown in the table below for each type of vehicle and different travel durations, correspond to the difference in time and cost between the two alternatives. The values used for the SP surveys are obtained by comparing the future costs associated to the project route (considering toll costs, and other operational costs) and costs that are reasonable for the routes that may present a competition for the project. The following table presents the difference in time and costs for each design. The deltas refer to a difference in time and cost between the trip done in the new route, and the trip done in the current route. The new alternative shows to be faster and at a larger cost. Steer Davies Gleave Page 196 of 259

97 Traffic study for project Conexión Antioquia - Bolívar Table 11-9: Difference between time and cost Light vehicles Design Trips shorter than 30 minutes Delta time (min) Delta cost ($) Trips between 3 30 minutes and 1 hour Delta time (min) Delta cost ($) Trips between 1 and 3 hours Delta time (min) Delta cost ($) Trips between 2 and 5 hours Delta time (min) Delta cost ($) Trips between 5 and 7 hours Delta time (min) Delta cost ($) Trips longer than 7 hours Delta time (min) Delta cost ($) 1 (5) 800 (10) 1,500 (15) 2,000 (30) 2,900 (45) 4,500 (60) 6,000 2 (10) 800 (20) 1,500 (30) 2,000 (45) 2,900 (60) 4,500 (90) 6,000 3 (15) - (25) 1,500 (40) 2,000 (60) 2,900 (90) 4,500 (120) 6,000 4 (5) 1,400 (10) 2,800 (15) 4,500 (30) 8,500 (45) 9,500 (60) 14,500 5 (10) 1,400 (20) 2,800 (30) 4,500 (45) 8,500 (60) 9,500 (90) 14,500 6 (15) 1,400 (25) 2,800 (40) 4,500 (60) 8,500 (90) 9,500 (120) 14,500 7 (5) 2,000 (10) 4,000 (15) 6,500 (30) 12,000 (45) 17,000 (60) 24,000 8 (10) 2,000 (20) 4,000 (30) 6,500 (45) 12,000 (60) 17,000 (90) 24,000 9 (15) 2,000 (25) 4,000 (40) 6,500 (60) 12,000 (90) 17,000 (120) 24,000 Source: Steer Davies Gleave Table 11-10: Difference between time and cost Small trucks Trips between30 Trips shorter than Trips between 1 minutes and 1 30 minutes and 3 hours hour Design Delta time (min) Delta cost ($) Delta time (min) Delta cost ($) Delta time (min) Delta cost ($) Trips between 2 and 5 hours Delta time (min) Delta cost ($) Trips between 5 and 7 hours Delta time (min) Delta cost ($) Trips longer than 7 hours Delta time (min) Delta cost ($) 1-5 $ 1, $ 2, $ 3, $ 4, $ 6, $ 8, $ 1, $ 2, $ 3, $ 4, $ 6, $ 8, $ 1, $ 2, $ 3, $ 4, $ 6, $ 8, $ 2, $ 4, $ 7, $ 13, $ 18, $ 25, $ 2, $ 4, $ 7, $ 13, $ 18, $ 25, $ 2, $ 4, $ 7, $ 13, $ 18, $ 25, $ 3, $ 6, $ 10, $ 21, $ 30, $ 45, $ 3, $ 6, $ 10, $ 21, $ 30, $ 45, $ 3, $ 6, $ 10, $ 21, $ 30, $ 45,000 Source: Steer Davies Gleave Steer Davies Gleave Page 197 of 259

98 Traffic study for project Conexión Antioquia - Bolívar Table 11-11: Difference between time and cost Large trucks Design Trips shorter than 30 minutes Delta time (min) Delta cost ($) Trips between 3 30 minutes and 1 hour Delta time (min) Delta cost ($) Trips between 1 and 3 hours Delta time (min) Delta cost ($) Trips between 2 and 5 hours Delta time (min) Delta time (min) Trips between 5 and 7 hours Delta cost ($) Delta time (min) Trips longer than 7 hours Delta cost ($) Delta time (min) 1-5 $ 1, $ 2, $ 4, $ 6, $ 9, $ 12, $ 1, $ 2, $ 4, $ 6, $ 9, $ 12, $ 1, $ 2, $ 4, $ 6, $ 9, $ 12, $ 3, $ 7, $ 9, $ 17, $ 20, $ 35, $ 3, $ 7, $ 9, $ 17, $ 20, $ 35, $ 3, $ 7, $ 9, $ 17, $ 20, $ 35, $ 5, $ 11, $ 15, $ 30, $ 45, $ 60, $ 5, $ 11, $ 15, $ 30, $ 45, $ 60, $ 5, $ 11, $ 15, $ 30, $ 45, $ 60,000 Source: Steer Davies Gleave The surveyed participants only see one version of the SP survey (9 cards each one) depending on the type of vehicle they drive and the travel time. It is important to highlight that the sequence in which the cards were presented varied for each participant, this way the potential biases related to learning and weariness during the exercise were minimized. Survey execution The survey was programmed on the software Sawtooth, which works together with the website Support from surveyors was provided in the different location points where current and potential users of the project were interviewed. This was necessary for heavy vehicles. Theory A type of demand model that is widely used in transport planning is the logit model. This type of model predicts the probability with which an individual chooses an alternative, considering the characteristics of the individual and the characteristics of the alternative. It is necessary to build a utility function that measures the satisfaction associated to each combination of alternative. This utility function is composed of two elements: i. A deterministic element related to observable and measurable attributes (for example, travel time, toll costs, etc.). ii. A random element that includes all other attributes that the modeler cannot know or measure (for example, fear of flying). Below is the utility function used: U ni V ni in Steer Davies Gleave Page 198 of 259

99 Traffic study for project Conexión Antioquia - Bolívar Where is a deterministic or measurable utility associated to mode, individual and corresponds to the random error associated to mode and individual, in other words, all variables known to the modeller. In general, is composed by characteristics from alternative and individual. In the estimated models for the project, presents a structure as the one described below: V ni x x i i n n In which is the array of characteristics from the alternative (for example, travel time, etc.), is the array of characteristics from the individual (for example, reason for travel, income, etc.) and and describe the effects of the alternative and individual perceived characteristics of utility or satisfaction. The Multinomial Logistic model (MNL) is used to model simple structures, in which all the errors come from equally distributed Gumbel functions. The general form of the Multinomial model is presented below: Vni P ni A A j e e n Vnj Where is the probability that the individual chooses alternative from the alternative array and where is the deterministic utility (for example, satisfaction) of alternative for each individual. Results from selection model for light vehicles The demand model that allows differentiating the value of time range for different segments assumes a division of the data according to their purpose. These were the purposes, or reason for travel, that were used in this model for light vehicles: Work (business) Tourism Other (recreation shopping, study, medical visits, other) The information below shows the logit model with these parameters. Table 11-12: Model with the cost parameter, segmented by trip purpose Light vehicle Purpose Coefficient Standard Error Test t Confidence interval 95% Cost Time-work Time-tourism Time-others Observations: 4,156 pseudo individuals Source: Steer Davies Gleave All variables in the model were significant and show: Coefficient sign: o Negative sign for cost and time: This shows that the utility is negative when these variables increase. For people in general it is preferred to reduce these variables since time and money are limited resources. Z Values: o Z values: values over 1.96 indicate a reliability level of 95%. These values show that the parameters in the results are highly reliable. This level of reliability is related to the sample size and variation of the answers obtained. Steer Davies Gleave Page 199 of 259

100 Traffic study for project Conexión Antioquia - Bolívar A key parameter in the demand model is the value of time. This indicator, also known as Subjective Value of Time (SVT, expressed in minutes), is defined as the division of time over cost. The following table shows the value of one minute, obtained in the calibrated logit model for light vehicles: Table 11-13: Value of time (COP/minutes) segmented by trip purpose Light vehicle Segment $COP/minute Standard Error Z Value of time, working trip Value of time, tourism trip Value of time, other purposes trip Source: Steer Davies Gleave This implies that the following values of time of one hour are: Table 11-14: Value of time (COP/hour) segmented by trip purpose Light vehicle Segment $COP/hour Value of time, working trip $ 12,332 Value of time, tourism trip $ 16,535 Value of time, other purposes trip $ 14,231 Source: Steer Davies Gleave Selection model results for trucks The calibrated model is shown below, for trucks. They were grouped in two categories: 2 axle trucks Trucks with 3 or more axles The following tables show the calibrated models for both categories. Table 11-15: Choice model Small trucks Parameter Coefficient Standard Error Z Confidence interval 95% Cost E Time Obs: 1,075 pseudo individuals Source: Steer Davies Gleave Table 11-16: Choice model Large trucks Parameter Coefficient Standard Error Z Confidence interval 95% Cost E Time Obs: 891 pseudo individuals Source: Steer Davies Gleave Like the calibrated model for light vehicles, these models are statistically significant at 95%, with the appropriate signs. Steer Davies Gleave Page 200 of 259

101 Traffic study for project Conexión Antioquia - Bolívar Applying time over cost, there is the value of time in minutes, which allows obtaining the value of time for one hour. Table 11-17: Value of time Trucks (COP/hour) Category camión $/hour z Small truck $18,560 33,37 Large truck $33,692 20,15 Source: Steer Davies Gleave Steer Davies Gleave Page 201 of 259

102 Traffic study for project Conexión Antioquia - Bolívar ANNEX F: OPINION SURVEYS RESULTS Sample and trip characteristics for light vehicles The following statistics correspond to light vehicles, which totaled 357 of the total surveys. Gender, occupation, age, and other characteristics were asked in the survey. Table 11-18: User characteristics Light vehicles Social strata Income Occupation Source: Steer Davies Gleave The following aspects are summarized from the figures: 90% of the respondents are male, 10% are women. 37% are between the ages of 31 and 40 years old. Most of the respondents are in socioeconomic levels 2 and 3, with 38% and 33% respectively. Regarding occupation, 35% are employed and 49% are self-employed. Regarding monthly incomes, 23% were in the range of $1,000,000 and $1,500,000 or between $2,000,000 y $3,000,000. One of the questions was regarding the trip purpose. The following figures show that the main purpose for the trip was work with 40% of the surveys, followed by tourism with 33% Steer Davies Gleave Page 202 of 259

103 Traffic study for project Conexión Antioquia - Bolívar Figure 11.9: Trip characteristics Light vehicles Trip purpose Trip frequency Source: Steer Davies Gleave Regarding travel frequency, 38% declare that their trip is occasional, while 17% say they that they travel once a week. Sample characteristics for trucks The following statistics correspond only to trucks (small and large), totaling 355 surveys. Figure 11.10: User characteristics Trucks Occupation Trip frequency Source: Steer Davies Gleave The following can be summarized: Most of them were employed with 73%, and 22 % were independent workers. Only 4% were contractors and less than 1% were retired or students. 40% make the trip at least once a week, followed by 21% who make the trip more than once a week. Steer Davies Gleave Page 203 of 259

104 Traffic study for project Conexión Antioquia - Bolívar Respondent characterization for other modes of transport Opinion surveys were applied for different modes of transport other than light vehicles and trucks, such as: Motorcycle Bus Taxi Mototaxi (motorbike taxi) Other Across these modes, 39% stated their trip was for work, 27% for home reasons. These results are shown in the following figures. Figure 11.1: User characteristics Other modes Trip purpose Trip frequency Source: Steer Davies Gleave Regarding travel frequency, 31% state that they make this trip occasionally, while 26% state that they make this trip daily. Analysis and results Figure 11.2: Trips responses Light vehicles Segment 1: Caucasia - Planeta Rica Segment Planeta Rica - Montería Steer Davies Gleave Page 204 of 259

105 Traffic study for project Conexión Antioquia - Bolívar Segment 3: Montería - Coveñas Segment 4: Coveñas - Tolú Segment 5: Tolú - Cruz del Viso Source: Steer Davies Gleave The figures above show the following behavior for each segment: Segment 1 Caucasia - Planeta Rica: if there is a 15% increase in tariffs, 94% of the users would continue to make the trip. When raising the tariffs 30% and 50%, 90% of the users would still make the trip in the same way. Segment 2 Planeta Rica - Montería: In all scenarios the tariff increase would not affect if the users would continue to make the same trips, showing that the demand is inelastic to the toll price. Segment 3 Montería - Coveñas: Almost all the users in light vehicles are willing to continue with their trip no matter what the tariff raise is. Only 4% of the users, with a 50% raise, would stop making the trip. This shows that in comparison with the area between Coveñas and Crus del Viso, there is higher economic capabilities in this area. Segment 4 Coveñas - Tolú: with a 15% increase in tariffs, 53% of the users continue to make the same trip. However, this percentage drops to 23% when there is a 30% increase and later to 8% when there is a 50% increase in toll prices. Segment 5 Tolú - Cruz del Viso: Regardless of the tariff raise, 91% of the users would still continue to make the same trip and only 9% would stop making it. Steer Davies Gleave Page 205 of 259

106 Traffic study for project Conexión Antioquia - Bolívar The same analysis was made for reason for travel and travel frequency for all segments. The results are shown in Annex D Small truck analysis and results Figure 11.3: Trips responses Small trucks Segment 1: Caucasia - Planeta Rica Segment 2: Planeta Rica - Montería Segment 3: Montería - Coveñas Segment 4: Coveñas - Tolú Segment 5: Tolú - Cruz del Viso Source: Steer Davies Gleave Steer Davies Gleave Page 206 of 259

107 Traffic study for project Conexión Antioquia - Bolívar Segment 1 Caucasia - Planeta Rica: Regardless of the tariff raise, 95% of the users would continue to make the trip Segment 2 Planeta Rica - Montería: 100% of the users are willing to continue their trip regardless of the tariff raise analyzed. Segment 3 Montería - Coveñas: Like in segment 1 and 2, almost of the sample is willing to continue with their trip. Only 7% of the users would stop making the trip with a 50% increase in toll prices. Segment 4 Coveñas - Tolú: 58% of the users would continue to make the trip with a 15% increase in tariffs. The number of users which continue to make the trip drops to 47% in the following scenarios. Segment 5 Tolú - Cruz del Viso: Like in segments 1, 2 and 3, regardless of the toll tariff raise, most of the users continue to make the trip. This could be caused by the amount of long distance trips that travel through this segment. Small trucks, in general, are willing to continue with their trip regardless of the tariff increase. Thus, the projections were not affected. Steer Davies Gleave Page 207 of 259

108 Traffic study for project Conexión Antioquia - Bolívar Large truck analysis and results Figure 11.4: Trips responses Large trucks Segment 1: Caucasia - Planeta Rica Segment 2: Planeta Rica - Montería Segment 3: Montería - Coveñas Segment 4: Coveñas - Tolú Segment 5: Tolú - Cruz del Viso Source: Steer Davies Gleave Steer Davies Gleave Page 208 of 259

109 Traffic study for project Conexión Antioquia - Bolívar Segment 1 Caucasia - Planeta Rica: Regardless of the increase, more than 96% of the users are willing to continue with their trip. Segment 2 Planeta Rica - Montería and Tramo 3 Montería - Coveñas: The same behavior from segment 1 applies. Segment 4 Coveñas - Tolú: With a 15% increase, 89% of the users are still willing to make the trip. The decrease is also 89% with a 30% increase in tariffs. With a 50% increase in tariffs however the number of users still willing to make the trip decreases to 85%. Segment 5 Tolú - Cruz del Viso: Like in segments 1, 2 and 3, 95% of the users would continue to make the trip regardless of the tariff increase. Like in the case of small trucks, there is a low elasticity, and so the projections were not affected for categories IV-VII. Steer Davies Gleave Page 209 of 259

110 Traffic study for project Conexión Antioquia - Bolívar ANNEX G: HISTORICAL TRAFFIC ANALYSIS Tolú - Sincelejo corridor This corridor connects Tolú Viejo and Sincelejo. La Esperanza toll is located exiting Tolú Viejo towards Sincelejo, where it can record long distance trips up to Cartagena from the western central region of the country, and regional trips from the area of Sincelejo and its surroundings. The following graph shows the percentage annual change in traffic. Figure 11.11: Yearly percentage change of the AADT and Colombia s GDP La Esperanza Source: Steer Davies Gleave based on vehicular records of each toll station The traffic recorded at this toll plaza shows a similar behavior to the other toll plazas which have a significant share of long distance trips (Cedros, Mata de Caña and San Onofre). In the period of there is a decrease in large truck traffic, however the total traffic continued to increase at lower rates. The period of shows an overall fall of traffic, where the volume is reduced by more than 100 thousand vehicles. From that year, traffic continued to increase and in 2013 traffic levels returned to what it had before. The effect of the economic boom that the region had in that period is evident in the traffic behavior in this toll. Similarly, to Cedros, Mata de Caña and San Onofre, there is a decrease in traffic after the year 2015, possibly associated to the change of the economy in the country, as has been mentioned in this chapter. Planeta Rica - Sincelejo - Carmen de Bolívar corridor This corridor has been historically the main corridor between Antioquia and the central-western part of the country with the Caribbean coast. It is characterized for having both local trips and long distance trips. The following graphs show a percentage annual growth in traffic for Carimagua and Las Flores tolls. El Carmen toll was also studied, however, there is incomplete information for the period. Steer Davies Gleave Page 210 of 259

111 Traffic study for project Conexión Antioquia - Bolívar The modal proportion in Carimagua tends to be more equitable than the other tolls that have been analyzed so far. This is consistent with the type of traffic which is expected to use this corridor. That is, long distance traffic with destination of Barranquilla or Cartagena from the center of the country, and that there are no important urban centers nearby. Las Flores toll station has available traffic history information from 2009 to In the first years ( ) total traffic increased at rates that varied between 5% and 9%, and are associated to the economy of the country in that period. Between 2013 and 2014, traffic decreased, and then recovered in Traffic then continued to grow in Figure 11.12: Yearly percentage change of the AADT and Colombia s GDP Carimagua Source: Steer Davies Gleave based on vehicular records of each toll station Steer Davies Gleave Page 211 of 259

112 Traffic study for project Conexión Antioquia - Bolívar Figure 11.13: Yearly percentage change of the AADT and Colombia s GDP Las Flores Source: Steer Davies Gleave based on vehicular records of each toll station Exits to ports (Cartagena y Barranquilla) This category comprises the exit from the Study Area of the project towards the port of Cartagena, represented by the toll plaza Gambote, and the exit to the port of Barranquilla, represented by the toll plaza Calamar. The graphs are shown below. Steer Davies Gleave Page 212 of 259

113 Traffic study for project Conexión Antioquia - Bolívar Figure 11.14: Yearly percentage change of the AADT and Colombia s GDP Calamar Source: Steer Davies Gleave based on vehicular records of each toll station Figure 11.15: Yearly percentage change of the AADT and Colombia s GDP Gambote Source: Steer Davies Gleave based on vehicular records of each toll station It is important to notice that the flow going through Gambote is almost twice the flow that goes through Calamar. This is because at this location the closest port is Cartagena. As such the reason for the difference is that cargo will take this path coming from or to Antioquia. Calamar has large variations in its annual traffic growth, for small trucks. There is a decrease in flow in the periods and Steer Davies Gleave Page 213 of 259

114 Traffic study for project Conexión Antioquia - Bolívar On the other hand, Gambote shows more abrupt changes in large trucks. There are decreases in total traffic in period and By the end of 2010 there was an important growth in flow, especially in heavy trucks, since the torrential rain caused damage to the Eastern road and the Cordialidad road which had to be closed for repairs. Thus, the Cartagena-Cruz del Viso corridor was the only alternative for trips originating from north of Bolívar and Atlántico heading towards the southwestern region. There is a fluctuation in large truck traffic in comparison to the other vehicles that have a more stable behavior after ANNEX H: NETWORK MODEL DEVELOPMENT The network in the model represents the road network in the study area. It is important that it is sufficiently detailed so that the route choice of travelers can be reasonably assessed. The network used in this study is based on data collection undertaken by Steer Davies Gleave during this study. The network was further refined based on a review of available documents and maps. The network is formed by a set of links connected by nodes. The basic information necessary to describe the nodes is their location, specified using coordinates and obtained through digital cartography. The relevant link information used for this study are as follows Start and end nodes Length Valid transport modes Number of lanes per direction Flow delay functions Capacity Tolls The base year road network corresponds to the conditions as at November Future year network assumptions as detailed in subsection 5.8. Note that while the main routes are included, not all available roads are. This is because the models are a simplified version of reality and so some filtering of unreasonable routes is required. The following figure illustrates the road network considered in the base year. Steer Davies Gleave Page 214 of 259

115 Traffic study for project Conexión Antioquia - Bolívar Figure 11.16: Model network, 2015 Source: Steer Davies Gleave Steer Davies Gleave Page 215 of 259

116 Traffic study for project Conexión Antioquia - Bolívar Flow-delay functions The relation between traffic volumes and delay is included in the model as a way of representing the impact of congestion. The most commonly used functions are the BPR function, given below: Where the travel time through the link section is a function of the traffic volumes, free flow time ( ), capacity for the road, and parameters and that define the effect of congestion. For example, the following figure shows an example flow-delay functions. The red curve corresponds to the function used for the Chinú - Sahagún section. This section is in good condition and has an average free flow speed of approximately 70 km/h. The grey curve represents the function used for the segment Lorica - San Andrés de Sotavento. This segment has lower speeds for similar capacity levels. Figure 11.17: Flow delay functions Source: Steer Davies Gleave Flow-delay functions depend on the free flow speed, link length, traffic volume including flows related to nonmodelled modes such as public transport buses, number of lanes, together with theoretic and practical capacities of the road The capacity corresponds to the maximum traffic flow that can pass by any given point in one hour under regular conditions. The capacity values used for each link vary based on the type of road as shown in the following table. Table 11-19: Lane capacity for various types of roads Capacity (Vehicles/hour/lane) Type of road (ranges in equivalent vehicles) Multi-lane 1,500-1,800 Service road 900-1,200 Local road 800-1,000 Source: Steer Davies Gleave Steer Davies Gleave Page 216 of 259

117 Traffic study for project Conexión Antioquia - Bolívar Travel demand Origin Destination Matrices To represent the travel demand, a series of matrices are used. These matrices store the origin and destination of trips occurring in the study area, for a specific day and period of time. Such matrices are commonly referred to as origin destination matrices, or travel matrices. The development of the travel matrices for this project was discussed in chapter 3 of this report. Surveys were carried out on road corridors to understand the trips using key roads. In these surveys, the main analysis unit is the trip, which corresponds to an individual vehicle. The surveys carried out represent a sample of the respective locations. Thus, the expansion of the sample is done through a comparison between the origin-destination couple, and the number of vehicles counted. For example, if one of the points shows that the trips made between Pasto and Ipiales correspond to 10% of the surveys, then it is assumed that the surveyed trip is representative of 10% of the counted trips. The surveys were carried out in places that are considered representative of the traffic that uses the road for this project. That is, long distance traffic with an origin or destination in the center-south or north coast of the country. Thus, the matrix has several empty cells that represent mainly the short trips between municipalities that were note captured at the survey locations. Although these trips are not part of the potential demand for the project, they are relevant as they have an impact on congestion. In order to account for these short distance trips, an adjustment to the travel matrices was required. This adjustment was based on observed vehicle counts. New demand matrices were obtained for each of the modelling periods, by type of vehicle, and by reason for travel. Note that only light vehicles are segmented by reason for travel. Given the demand characteristics of the Study Area, it was decided to build a travel matrix for two time periods: Average hour for a weekday between 06:00 and 18:00 Average hour for a weekend between 06:00 and 18:00 For each time, an average volume was determined and the matrices were built for the following types of vehicles and types of users: Light vehicles (% weekday / % weekend) Work/business trips (47% / 29%) Tourism trips (5% / 17%) Other trips (48% / 54%) Small trucks Large trucks Thus, the origin-destination matrices, capturing both long and short distance trips, were obtained for 2015, the base year to be modelled. The table below presents the total trips per hour for the matrices used in the network model. Table 11-20: Demand matrices per hour Period Type of vehicle Total vehicles (trips/hour) Light vehicles 918 Weekday Small Trucks 316 Large Trucks 190 Steer Davies Gleave Page 217 of 259

118 Traffic study for project Conexión Antioquia - Bolívar Period Type of vehicle Total vehicles (trips/hour) Light vehicles 958 Weekend Small Trucks 226 Large Trucks 148 Source: Steer Davies Gleave Since public transport vehicle use does not behave in a similar way to private vehicles, public transport vehicles were not included in the model. However, these vehicles were accounted for as part of the calibration. In models used for these types of studies, it is not common practice to pre-load the public transport vehicles onto the network due to the following reasons: This study is based on an interurban network, on which there is no congestion. The pre-load of the network is commonly used to represent the effect of public transport vehicles on capacity, congestion and hence speed and travel times. As mentioned above however this is not relevant to this study. This study found that on average buses represent 8% of the total traffic. Therefore, it is not considered that they generate an important effect on the routes. Public transport routes are not defined by the congestion that a road may have, nor by the costs of different alternatives. Rather they are defined based on a desire to connect users. As such the route choice of public transport vehicles are not required to be explicitly modelled. Model Zoning The zoning used is focused on developing a strategic model at a regional level. This is the reason for which the area was divided into an aggregated system of 493 areas that represent the municipalities within the study area (Antioquia, Bolívar, Córdoba, among others) and areas that represent municipality groups for the rest of the country as shown in the figure below. Each of the areas are also represented by a centroid. The centroid is the location from which all the trips in the area are loaded. The centroids are also shown in the following figure. Steer Davies Gleave Page 218 of 259

119 Traffic study for project Conexión Antioquia - Bolívar Figure 11.18: Model Zoning (centroids) Source: Steer Davies Gleave Steer Davies Gleave Page 219 of 259

120 Traffic study for project Conexión Antioquia - Bolívar Decision parameters Toll stations The costs associated to charges in tolls for the use of the roads in the area of study are included in the model according to current locations and consistent with the information presented in the project summary. Values of time In Chapter 3 of this report, the analysis for stated preferences was discussed. These surveys were used to estimate the willingness to pay different tolls on the roads in and around the study area. As a result, three types of vehicle classes were determined. These vehicle classes are Light vehicles; Small trucks (2 axles); and Large trucks (more than 3 axles). The values of time estimated for each of these vehicle classes types were incorporated into the network model. The following table shows the values of time used in the model. Table 11-21: Values of time (COP 2015) Type of vehicle Segment Value of time (COP $/h) 1 Lightweight vehicles Work/Business $12,332 2 Lightweight vehicles Tourism/ other $16,535 3 Lightweight vehicles Other $14,231 5 Small trucks 2 axles $18,664 6 Large trucks More than 2 axles $33,904 Source: Steer Davies Gleave Operating costs The vehicle operating costs for all types of vehicles and roads were calculated using published values in the INVIAS booklet of Volúmenes de tránsito and costos de operación (Traffic volumes and operating costs ). These values are estimated with the use of the HDM4 software. The values of time calculated were indexed to 2015, the base year, based on the inflation reported by the National Department of Statistics. The values used are shown below. The values presented in the INVIAS booklet include the complete costs associated to a trip. For example, fuel, tires, maintenance, salaries, ect. It is assumed that, in terms of the route selection, vehicle users only perceive the fuel cost, which accounts for 43% of the total cost. This percentage was estimated based on the cost structure for road transport developed by ANIF. The following table summarizes the final operating costs included in the base scenarios for the network model. Table 11-22: Perceived operating costs (COP 2015) Light vehicles operating Small trucks Large trucks Segment costs (COP/km) operating costs (COP/km) operating costs (COP/km) Montería - Los Pantanos $1,558 Montería - Patio Bonito $1,565 Patio Bonito - Planeta Rica $1,565 Steer Davies Gleave Page 220 of 259

121 Traffic study for project Conexión Antioquia - Bolívar Segment Light vehicles operating costs (COP/km) Small trucks operating costs (COP/km) Large trucks operating costs (COP/km) Planeta Rica - La Apartada $1,565 Planeta Rica - La Ye $1,710 La Ye - Ciénaga de Oro $1,565 Ciénaga de Oro - Cereté $1,565 Cereté - Lorica $1,558 Lorica - San Andrés de Sotavento $1,806 San andres de sotavento - chinú $1,710 Chinú - Sahagún $1,580 Sahagún - La Ye $1,565 Cereté - Montería $1,565 Los Pantanos - Santa Lucia $1,694 Santa Lucía - Corregimiento Bajo Blanco Corregimiento Bajo Blanco - Moñitos $2, $1,694 Moñitos - San Bernardo del Viento $1,694 San Bernardo del Viento - Lorica $1,694 Tolú - San Onofre $2,297 Tolúviejo - Sincelejo $1,710 Tolú - Tolúviejo $1,694 Coveñas - Tolú $1,797 Lorica - San Antero $1,694 Tolúviejo - San Onofre $1,694 San Onofre - María La Baja $1,694 María La Baja - Cruz del Viso $1,694 Cruz del Viso - San Cayetano $1,565 Carmen de Bolívar - Carreto $1,580 Corozal - Carmen de Bolívar $1,580 Chinú - Corozal $1,580 Calamar - Carreto $1,687 Note: Inflation used for these estimates is reported by DANE Source: Steer Davies Gleave based on values from INVIAS 2013 Steer Davies Gleave Page 221 of 259

122 Traffic study for project Conexión Antioquia - Bolívar Model calibration By combining the various aspects discussed above, a traffic model was built for the base year (2015). This model was then calibrated to ensure that it can reasonably replicate current conditions. This task consists on comparing observed and modelled traffic volumes and speeds. This process involves the following key tasks. Verity that all the important links are accurately represented Ensure realistic flow-delay functions are used Confirm that realistic operating costs for different vehicles are assumed The user categories are sufficient to represent the travel behavior and that these categories have a value of time that is appropriate for use in the model Verify that the travel matrices used are realistic and their development has been audited and reviewed. The calibration process can be understood graphically with the following figure. Figure 11.19: Description of calibration process Source: Steer Davies Gleave Traffic modelling is carried out using an algorithm that looks to select the optimal route for each type of user. That is, to minimize the generalized travel cost given by the following equation. In order to predict the possible variability of travel times in the study area, a stochastic assignment is used for this project. This type of assignment assumes that all users choose the route that has the lowest individual perceived costs, but it includes lower variability in costs to represent the imperfect information that users have regarding costs and benefits. A key part of the calibration process is the comparison of modelled flows against observed values. These are presented in the next section. Comparison of modelled and observed vehicle flows A common way to quantify the extent to which a model replicates current conditions is to compare traffic counts to link flows. These can then be compared across the network as a whole using an R2 indicator defined below. Steer Davies Gleave Page 222 of 259

123 Traffic study for project Conexión Antioquia - Bolívar The R 2 coefficient indicates the variance of the model from observed traffic volumes. Differences on individual links can also be compared in relative terms. It is considered generally acceptable to aim for all differences to be within 20%. However, in the case of small levels of observed flows a percentage measure is not considered to be a good measure and so it is recommended to complete the analysis with other indicators, such a GEH. The GEH statistic, named after Geoffrey E. Havers, is an indicator that considers the order of magnitude of the values compared. It corresponds to an approximation of a statistic that follows the chi-square distribution. It is considered that the values between 0 and 5 indicate a good adjustment, between 5 and 10 are acceptable and larger than 10 require further investigation. The GEH indicator is recommended when large percentage differences on small values are irrelevant. Generally it is accepted that a traffic assignment model is calibrated satisfactorily if one or all of the following conditions are achieved 18 : At least 60% of the links have a GEH less than 5 95% of the links have a GEH less than 10 All the links have a GEH less than 12 The expected GEH for all screen line comparisons is less than 4 The following tables and graphs show a comparison between the observed and modelled flows. This comparison is provided for light vehicles, small trucks, and large trucks. These volumes were obtained through the fieldwork as described previously. In all the calibration periods and for all vehicular categories, the R 2 coefficient is larger than Additionally, all count sites have a GEH smaller than 2. In general, these indicators show that the network assignment model is satisfactorily calibrated. 18 Modelación de Demanda para Carreteras de Cuota. (2006) Transconsult, Secretaria de Comunicaciones y Transportes de México, Steer Davies Gleave Steer Davies Gleave Page 223 of 259

124 Traffic study for project Conexión Antioquia - Bolívar Figure 11.20: Comparison of vehicular flows modelled and counted Light vehicles Source: Steer Davies Gleave Table 11-23: Comparison of vehicular flows modelled and counted Light vehicles ID Location Direction M1 M1 M2 M2 M3 M3 M4 M4 M5 M5 M6 M6 M7 Weekday Observed (veh/h) Modelled (veh/h) Dif. % GEH Weekend Observed (veh/h) Modelled (veh/h) Dif. % Toll La Apartada N-S % % 0.2 Toll La Apartada S-N % % 0.1 Toll Purgatorio N-S % % 0.2 Toll Purgatorio S-N % % 0.3 Toll Cedros N-S % % 0.1 Toll Cedros S-N % % 0.2 Toll Carimagu N-S % % 0.2 a Toll Carimagu S-N % % 0.3 a Toll Mata de Caña N-S % % 0.3 Toll Mata de Caña S-N % % 0.5 Toll La Esperanza N-S % % 0.2 Toll La Esperanza S-N % % 0.1 Toll San Onofre N-S % % 0.2 GEH Steer Davies Gleave Page 224 of 259

125 Traffic study for project Conexión Antioquia - Bolívar ID Location Direction M7 M8 M8 Toll San Onofre Toll El Carmen Toll El Carmen Weekday Observed (veh/h) Modelled (veh/h) Dif. % GEH Weekend Observed (veh/h) Modelled (veh/h) Dif. % S-N % % 0.5 N-S % % 0.1 S-N % % 0 M9 Sahagún N-S % % 0.8 M9 Sahagún S-N % % 0.7 M10 M10 M11 M11 M12 M12 San Antero San Antero Cruz Del Viso Cruz Del Viso Campoal egre Campoal egre N-S % % 1.2 S-N % % 0.1 N-S % % 0.3 S-N % % 0.9 N-S % % 0.8 S-N % % 0.4 Note: Obs.=observed, Mod.= modelle, Dif. % = percentual difference Source: Steer Davies Gleave Figure 11.21: Comparison of vehicular flows modelled and counted Small trucks GEH Source: Steer Davies Gleave Steer Davies Gleave Page 225 of 259

126 Traffic study for project Conexión Antioquia - Bolívar Table 11-24: Comparison of vehicular flows modelled and counted Small trucks ID Location Direction M1 M1 M2 M2 M3 M3 M4 M4 M5 M5 M6 M6 M7 M7 M8 M8 Toll La Apartada Toll La Apartada Toll Purgatori o Toll Purgatori o Toll Cedros Toll Cedros Toll Carimag ua Toll Carimag ua Toll Mata de Caña Toll Mata de Caña Toll La Esperanz a Toll La Esperanz a Toll San Onofre Toll San Onofre Toll El Carmen Toll El Carmen Weekday Observed (veh/h) Modelled (veh/h) Dif. % GEH Weekend Observed (veh/h) Modelled (veh/h) Dif. % N-S % % 0.1 S-N % % 0 N-S % % 0.1 S-N % % 0.2 N-S % % 0.2 S-N % % 0.1 N-S % % S-N % % 0.1 N-S % % 0 S-N % % 0.1 N-S % % S-N % % 0.1 N-S % % 0.2 S-N % % N-S % % 0 S-N % GEH % 0.4 M9 Sahagún N-S % % 0.1 Steer Davies Gleave Page 226 of 259

127 Traffic study for project Conexión Antioquia - Bolívar M9 Sahagún S-N M10 M10 M11 M11 M12 M12 San Antero San Antero Cruz Viso Cruz Viso Del Del Campoal egre Campoal egre % Note: Obs.=observed, Mod.= modelled, Dif. % = percentage difference Source: Steer Davies Gleave Figure 11.22: Comparison of vehicular flows modelled and counted Large trucks % 0.3 N-S % % 0.2 S-N % % 0.1 N-S % S-N % N-S % % % % 0 S-N % % Source: Steer Davies Gleave Table 11-25: Comparison of vehicular flows modelled and counted Large trucks ID Location Direction M1 M1 Toll Apartada Toll Apartada La La Weekday Obs. (veh/h) Mod. (veh/h) Dif. % GEH Weekend Obs. (veh/h) Mod. (veh/h) Dif. % N-S % % 0.2 S-N % % 0.7 M2 Toll Purgatorio N-S % % 0.1 GEH Steer Davies Gleave Page 227 of 259

128 Traffic study for project Conexión Antioquia - Bolívar ID Location Direction Weekday Obs. (veh/h) Mod. (veh/h) Dif. % GEH Weekend Obs. (veh/h) Mod. (veh/h) Dif. % M2 Toll Purgatorio S-N % % 0.4 M3 Toll Cedros N-S % % 0 M3 Toll Cedros S-N % % 0.3 M4 M4 M5 M5 M6 M6 M7 M7 Toll Carimagua Toll Carimagua Toll Mata de Caña Toll Mata de Caña Toll La Esperanza Toll La Esperanza Toll San Onofre Toll San Onofre N-S % % 0 S-N % % 0.3 N-S % % 0.5 S-N % % 0.1 N-S % % 0.2 S-N % % 0 N-S % % 0.4 S-N % % 0.3 M8 Toll El Carmen N-S % % 0.7 M8 Toll El Carmen S-N % % 0.6 M9 Sahagún N-S % % 0.6 M9 Sahagún S-N % % 0.5 M10 San Antero N-S % % 0 M10 San Antero S-N % % 0.2 M11 Cruz Del Viso N-S % % 1.7 M11 Cruz Del Viso S-N % % 0.9 M12 Campoalegre N-S % % 0.4 M12 Campoalegre S-N % % 0.1 Note: Obs.=observed, Mod.= modelled, Dif. % = percentage difference Source: Steer Davies Gleave GEH Speed calibration Along with the traffic flows, the calibration process includes a comparison between the modelled and observed speeds. In general, the target values for the calibration of travel times are +/- 15% in each corridor. As with traffic counts, the GEH can also be used and is estimated for each segment of interest The following table summarizes the calculated values for the main corridors. The calibrated speed is the average speed observed for light vehicles for weekdays and weekend days, in some segments, this value is slightly different from the observed information. Steer Davies Gleave Page 228 of 259

129 Traffic study for project Conexión Antioquia - Bolívar Table 11-26: Speed calibration Segment Calibration speed (km/h) Modelled speed (km/h) Difference % GEH Montería - Los Pantanos % 0.5 Montería - Patio Bonito % 0 Patio Bonito - Planeta Rica % 0.7 Planeta Rica - La Apartada % 0 Planeta Rica - La Ye % 0 La Ye - Ciénaga de Oro % 0.1 Ciénaga de Oro - Cereté % 0.5 Cereté - Lorica % 0.6 Lorica - San Andrés de Sotavento % 1.1 San andres de sotavento - chinú % 0.1 Chinú - Sahagún % 0.9 Sahagún - La Ye % 0.9 Cereté - Montería } % 0.1 Los Pantanos - Santa Lucia % 0.4 Santa Lucía - Corregimiento Bajo Blanco % 0.5 Corregimiento Bajo Blanco - Moñitos % 0.1 Moñitos - San Bernardo del Viento % 0.5 San Bernardo del Viento - Lorica % 0.2 Tolú - San Onofre % 0.4 Tolúviejo - Sincelejo % 0.1 Tolú - Tolúviejo % 0 Steer Davies Gleave Page 229 of 259

130 Traffic study for project Conexión Antioquia - Bolívar Segment Calibration speed (km/h) Modelled speed (km/h) Difference % GEH Coveñas - Tolú % 0.4 Lorica - San Antero % 0.1 Tolúviejo - San Onofre % 0.3 San Onofre - María La Baja % 0.3 María La Baja - Cruz del Viso % 0.1 Cruz del Viso - San Cayetano % 0.1 Carmen de Bolívar - Carreto % 0.5 Corozal - Carmen de Bolívar % 0.4 Chinú - Corozal % 0.5 Calamar - Carreto % 0.2 Source: Steer Davies Gleave All of the routes above have a GEH lower than 2, showing that the speeds used in the model closely replicate observed conditions. Steer Davies Gleave Page 230 of 259

131 Traffic study for project Conexión Antioquia - Bolívar ANNEX I: ANNUALISATION FACTORS The following table shows the expansion factors to go from average hour to daily demand. This is shown for both weekday and weekend Table 11-27: Expansion factors, modelling hour to day (weekday and weekend) Steer Davies Gleave Page 231 of 259

132 Traffic study for project Conexión Antioquia - Bolívar Expansion factor: model hour- weekday Expansion factor: model hour- weekend day Station Direction Light vehicles Buses Light vehicles Buses Small trucks Large trucks Small trucks Large trucks M1 Toll La Apartada Montería - Medellín M1 Toll La Apartada Montería - Medellín M2 Toll Purgatorio Montería - Planeta Rica M2 Toll Purgatorio Planeta Rica - Montería M3 Toll Cedros Arboletes - Montería M3 Toll Cedros Montería - Arboletes M5 Toll Mata de Caña Lorica - Cerete M5 Toll Mata de Caña Cerete - Lorica M7 Toll San Onofre Cartagena - Montería M7 Toll San Onofre Montería - Cartagena Steer Davies Gleave Page 232 of 259

133 Traffic study for project Conexión Antioquia - Bolívar M10 San Antero Coveñas - Lorica M10 San Antero Lorica - Coveñas Source: Steer Davies Gleave Steer Davies Gleave Page 233 of 259

134 Traffic study for project Conexión Antioquia - Bolívar The values presented in the previous table indicate the spread of traffic across the day. That is, given that the modelled hour represents an average hourly flow from 6am to 6pm we know the following. An expansion factor of 12 implies that there is no traffic between 6 pm and 6 am A factor of 24 implies that traffic at night is the same as the one during the day If the expansion factor is higher than 24, night traffic is higher than the during the day. The following table shows the weekday to weekly expansion factors. For weekends days, expansion factors are calculated to be double the average weekend day traffic. Table 11-28: Expansion factors from weekday hour to week Station Direction Expansion factor: weekday-week Light vehicles Buses Small trucks Large trucks M1 Toll La Apartada Montería - Medellín M1 Toll La Apartada Medellín - Montería M2 Toll Purgatorio Montería - Planeta Rica M2 Toll Purgatorio Planeta Rica - Montería M3 Toll Cedros Arboletes - Montería M3 Toll Cedros Montería - Arboletes M4 Toll Carimagua Sahagún - Planeta Rica M4 Toll Carimagua Planeta Rica - Sahagún M5 Toll Mata de Caña Lorica - Cerete M5 Toll Mata de Caña Cerete - Lorica M7 Toll San Onofre Cartagena - Montería M7 Toll San Onofre Montería - Cartagena Source: Steer Davies Gleave The following table presents the weekly to monthly expansion factors. Steer Davies Gleave Page 234 of 259

135 Traffic study for project Conexión Antioquia - Bolívar Table 11-29: Weekly to monthly expansion factors Expansion factor: week to month Toll Light vehicles Buses Small trucks Large trucks La Apartada Mata de Caña San Onofre Source: Steer Davies Gleave We expect the factors to be close to 4.3 the number of weeks in November as it is normal for traffic to be evenly distributed during the month. A lower value, as shown above, indicates that the traffic in the week of data collection is higher than the average weekly traffic. Lastly, the following table shows the monthly to annual expansion factors. These are estimated are based on monthly historic information at the toll plazas in the study area. This was available from 2004 to Table 11-30: Expansion factors from month to year Station Toll Light vehicles Buses Small trucks Large trucks M1 Toll La Apartada M2 Toll Purgatorio M3 Toll Cedros M5 Toll Mata de Caña M7 Toll San Onofre M10 San Antero Source: Steer Davies Gleave In the table, above, factors higher than 12 indicate that the traffic for November is comparatively lower than the average monthly traffic for the corresponding vehicle category. Conversely when factors are lower than 12, this means that the traffic in November is higher to the average monthly traffic. Steer Davies Gleave Page 235 of 259

136 Traffic study for project Conexión Antioquia - Bolívar ANNEX J: TRAFFIC PROJECTIONS FOR THE SENSITIVITY TESTS The results are presented in total ADDT for each toll plaza Table 11-31: Traffic projections, sensitivity tests La Apartada Year Base scenario SI SII SIII SIV SV ,086 4,086 4,086 4,086 4,150 4,034 4,225 4,225 4,225 4,225 4,312 4,153 4,373 4,373 4,373 4,373 4,489 4,282 4,517 4,454 4,454 4,517 4,659 4,405 4,676 4,611 4,611 4,676 4,846 4,538 4,839 4,773 4,773 4,839 5,039 4,677 5,011 4,943 4,943 5,011 5,241 4,820 5,117 5,117 4,664 5,203 5,377 4,898 5,277 5,276 4,807 5,365 5,571 5,028 5,441 5,440 4,955 5,531 5,772 5,160 5,611 5,610 5,107 5,703 5,981 5,297 5,786 5,784 5,264 5,880 6,197 5,436 5,967 5,964 5,426 6,063 6,421 5,580 6,146 6,142 5,586 6,244 6,644 5,722 6,331 6,325 5,751 6,431 6,875 5,867 6,522 6,513 5,921 6,623 7,115 6,016 6,718 6,708 6,095 6,821 7,363 6,169 6,921 6,908 6,276 7,025 7,619 6,325 7,059 7,046 6,401 7,166 7,810 6,420 7,200 7,187 6,529 7,309 8,005 6,517 7,344 7,331 6,660 7,455 8,205 6,614 7,491 7,477 6,793 7,605 8,410 6,714 7,641 7,627 6,929 7,757 8,620 6,814 7,794 7,779 7,067 7,912 8,836 6,917 Steer Davies Gleave Page 236 of 259

137 Traffic study for project Conexión Antioquia - Bolívar Year Base scenario SI SII SIII SIV SV ,950 7,935 7,209 8,070 9,057 7,020 8,109 8,094 7,353 8,231 9,283 7,126 8,271 8,255 7,500 8,396 9,515 7,232 8,436 8,421 7,650 8,564 9,753 7,341 8,605 8,589 7,803 8,735 9,997 7,451 8,777 8,761 7,959 8,910 10,247 7,563 8,952 8,936 8,118 9,088 10,503 7,676 9,132 9,115 8,280 9,270 10,766 7,791 9,314 9,297 8,446 9,455 11,035 7,908 9,500 9,483 8,615 9,644 11,311 8,027 9,690 9,673 8,787 9,837 11,594 8,147 9,884 9,866 8,963 10,034 11,883 8,270 10,082 10,063 9,142 10,235 12,180 8,394 10,284 10,265 9,325 10,439 12,485 8,519 Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Table 11-32: Percentage change between sensitivities and base case La Apartada Year SI SII SIII SIV SV % 0% 0% 2% -1% % 0% 0% 2% -2% % 0% 0% 3% -2% % -1% 0% 3% -2% % -1% 0% 4% -3% % -1% 0% 4% -3% % -1% 0% 5% -4% % -9% 2% 5% -4% % -9% 2% 6% -5% % -9% 2% 6% -5% % -9% 2% 7% -6% % -9% 2% 7% -6% % -9% 2% 8% -6% % -9% 2% 8% -7% Steer Davies Gleave Page 237 of 259

138 Traffic study for project Conexión Antioquia - Bolívar Year SI SII SIII SIV SV % -9% 2% 9% -7% % -9% 2% 9% -8% % -9% 2% 10% -8% % -9% 2% 10% -9% % -9% 2% 11% -9% % -9% 2% 11% -9% % -9% 2% 12% -10% % -9% 2% 12% -10% % -9% 2% 13% -11% % -9% 2% 13% -11% % -9% 2% 14% -12% % -9% 2% 14% -12% % -9% 2% 15% -13% % -9% 2% 16% -13% % -9% 2% 16% -13% % -9% 2% 17% -14% % -9% 2% 17% -14% % -9% 2% 18% -15% % -9% 2% 18% -15% % -9% 2% 19% -16% % -9% 2% 20% -16% % -9% 2% 20% -16% % -9% 2% 21% -17% % -9% 2% 21% -17% Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 238 of 259

139 Traffic study for project Conexión Antioquia - Bolívar Tabla 11-33: Traffic projections, sensitivity tests Los Manguitos Year Base scenario SI SII SIII SIV SV ,694 5,694 5,694 5,694 5,782 5,623 5,886 5,886 5,886 5,886 6,007 5,791 6,093 6,093 6,093 6,093 6,255 5,972 6,299 6,235 6,235 6,299 6,497 6,148 6,521 6,456 6,456 6,521 6,758 6,335 6,749 6,682 6,682 6,749 7,027 6,530 6,990 6,921 6,921 6,990 7,308 6,731 7,162 7,165 6,713 7,262 7,524 6,866 7,386 7,389 6,921 7,488 7,796 7,048 7,617 7,619 7,135 7,721 8,077 7,235 7,855 7,856 7,355 7,961 8,369 7,427 8,100 8,102 7,583 8,209 8,672 7,624 8,354 8,355 7,818 8,465 8,985 7,826 8,606 8,605 8,050 8,719 9,298 8,026 8,866 8,863 8,290 8,981 9,623 8,231 9,134 9,129 8,538 9,251 9,959 8,441 9,410 9,402 8,792 9,530 10,306 8,657 9,694 9,685 9,055 9,817 10,667 8,878 9,888 9,878 9,236 10,013 10,933 9,011 10,086 10,076 9,420 10,213 11,207 9,146 10,288 10,277 9,609 10,418 11,487 9,283 10,494 10,483 9,801 10,626 11,774 9,423 10,703 10,692 9,997 10,838 12,068 9,564 10,917 10,906 10,197 11,055 12,370 9,708 11,136 11,124 10,401 11,276 12,679 9,853 11,359 11,347 10,609 11,502 12,996 10,001 Steer Davies Gleave Page 239 of 259

140 Traffic study for project Conexión Antioquia - Bolívar Year Base scenario SI SII SIII SIV SV ,586 11,574 10,821 11,732 13,321 10,151 11,817 11,805 11,037 11,966 13,654 10,303 12,054 12,041 11,258 12,206 13,995 10,458 12,295 12,282 11,483 12,450 14,345 10,615 12,541 12,528 11,713 12,699 14,704 10,774 12,792 12,779 11,947 12,953 15,072 10,935 13,047 13,034 12,186 13,212 15,448 11,099 13,308 13,295 12,430 13,476 15,835 11,266 13,574 13,561 12,679 13,746 16,230 11,435 13,846 13,832 12,932 14,021 16,636 11,606 14,123 14,109 13,191 14,301 17,052 11,781 14,405 14,391 13,455 14,587 17,478 11,957 Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Table 11-34: Percentage change between sensitivities and base case Los Manguitos Year SI SII SIII SIV SV % 0% 0% 2% -1% % 0% 0% 2% -2% % 0% 0% 3% -2% % -1% 0% 3% -2% % -1% 0% 4% -3% % -1% 0% 4% -3% % -1% 0% 5% -4% % -6% 1% 5% -4% % -6% 1% 6% -5% % -6% 1% 6% -5% % -6% 1% 7% -5% % -6% 1% 7% -6% % -6% 1% 8% -6% % -6% 1% 8% -7% % -6% 1% 9% -7% % -7% 1% 9% -8% % -7% 1% 10% -8% % -7% 1% 10% -8% Steer Davies Gleave Page 240 of 259

141 Traffic study for project Conexión Antioquia - Bolívar Year SI SII SIII SIV SV % -7% 1% 11% -9% % -7% 1% 11% -9% % -7% 1% 12% -10% % -7% 1% 12% -10% % -7% 1% 13% -11% % -7% 1% 13% -11% % -7% 1% 14% -12% % -7% 1% 14% -12% % -7% 1% 15% -12% % -7% 1% 16% -13% % -7% 1% 16% -13% % -7% 1% 17% -14% % -7% 1% 17% -14% % -7% 1% 18% -15% % -7% 1% 18% -15% % -7% 1% 19% -15% % -7% 1% 20% -16% % -7% 1% 20% -16% % -7% 1% 21% -17% % -7% 1% 21% -17% Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 241 of 259

142 Traffic study for project Conexión Antioquia - Bolívar Table 11-35: Traffic projections, sensitivity tests Los Cedros Year Base scenario SI SII SIII SIV SV ,982 1,982 1,982 1,982 2,008 1,956 2,045 2,045 2,045 2,045 2,081 2,010 2,116 2,116 2,116 2,116 2,163 2,071 2,213 2,207 2,207 2,213 2,272 2,155 2,291 2,284 2,284 2,291 2,363 2,221 2,371 2,363 2,363 2,371 2,456 2,289 2,455 2,447 2,447 2,455 2,553 2,360 2,566 2,533 2,495 2,541 2,682 2,456 2,647 2,612 2,572 2,620 2,779 2,521 2,730 2,694 2,652 2,702 2,880 2,588 2,815 2,778 2,735 2,787 2,985 2,657 2,904 2,865 2,820 2,874 3,093 2,728 2,995 2,955 2,908 2,964 3,206 2,801 3,087 3,046 2,997 3,055 3,321 2,873 3,182 3,141 3,089 3,149 3,440 2,948 3,281 3,239 3,184 3,246 3,564 3,024 3,382 3,340 3,283 3,346 3,692 3,103 3,487 3,444 3,384 3,449 3,825 3,183 3,556 3,513 3,452 3,518 3,920 3,231 3,628 3,583 3,521 3,588 4,018 3,279 3,700 3,654 3,591 3,660 4,119 3,328 3,774 3,728 3,663 3,733 4,222 3,378 3,850 3,802 3,736 3,808 4,327 3,429 3,927 3,878 3,811 3,884 4,436 3,480 4,005 3,956 3,887 3,962 4,546 3,533 Steer Davies Gleave Page 242 of 259

143 Traffic study for project Conexión Antioquia - Bolívar Year Base scenario SI SII SIII SIV SV 4,085 4,035 3,965 4,041 4,660 3, ,167 4,116 4,044 4,122 4,777 3,639 4,250 4,198 4,125 4,204 4,896 3,694 4,335 4,282 4,207 4,288 5,018 3,749 4,422 4,367 4,292 4,374 5,144 3,806 4,510 4,455 4,377 4,462 5,272 3,863 4,601 4,544 4,465 4,551 5,404 3,921 4,693 4,635 4,554 4,642 5,539 3,979 4,786 4,727 4,645 4,735 5,678 4,039 4,882 4,822 4,738 4,829 5,820 4,100 4,980 4,918 4,833 4,926 5,965 4,161 5,079 5,017 4,930 5,025 6,114 4, ,181 5,117 5,028 5,125 6,267 4,287 Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Table 11-36: Percentage change between sensitivities and base case Los Cedros Year SI SII SIII SIV SV % 0% 0% 1% -1% % 0% 0% 2% -2% % 0% 0% 2% -2% % 0% 0% 3% -3% % 0% 0% 3% -3% % 0% 0% 4% -3% % 0% 0% 4% -4% % -3% -1% 5% -4% % -3% -1% 5% -5% % -3% -1% 5% -5% % -3% -1% 6% -6% % -3% -1% 7% -6% % -3% -1% 7% -6% % -3% -1% 8% -7% % -3% -1% 8% -7% % -3% -1% 9% -8% % -3% -1% 9% -8% Steer Davies Gleave Page 243 of 259

144 Traffic study for project Conexión Antioquia - Bolívar Year SI SII SIII SIV SV % -3% -1% 10% -9% % -3% -1% 10% -9% % -3% -1% 11% -10% % -3% -1% 11% -10% % -3% -1% 12% -10% % -3% -1% 12% -11% % -3% -1% 13% -11% % -3% -1% 14% -12% % -3% -1% 14% -12% % -3% -1% 15% -13% % -3% -1% 15% -13% % -3% -1% 16% -14% % -3% -1% 16% -14% % -3% -1% 17% -14% % -3% -1% 17% -15% % -3% -1% 18% -15% % -3% -1% 19% -16% % -3% -1% 19% -16% % -3% -1% 20% -16% % -3% -1% 20% -17% % -3% -1% 21% -17% Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 244 of 259

145 Traffic study for project Conexión Antioquia - Bolívar Table 11-37: Traffic projections, sensitivity tests Purgatorio Year Base scenario SI SII SIII SIV SV ,014 5,014 5,014 5,014 5,094 4,960 5,162 5,162 5,162 5,162 5,274 5,092 5,343 5,343 5,343 5,343 5,494 5,257 5,535 5,532 5,532 5,535 5,719 5,427 5,728 5,724 5,724 5,728 5,948 5,591 5,937 5,923 5,923 5,937 6,194 5,771 6,150 6,135 6,135 6,150 6,441 5,950 6,377 6,352 6,239 6,405 6,709 6,144 6,576 6,550 6,434 6,605 6,950 6,308 6,782 6,755 6,634 6,812 7,200 6,476 6,993 6,966 6,841 7,025 7,459 6,649 7,212 7,184 7,054 7,245 7,728 6,826 7,437 7,408 7,274 7,471 8,007 7,008 7,663 7,633 7,494 7,699 8,287 7,188 7,895 7,865 7,721 7,933 8,578 7,373 8,134 8,103 7,955 8,174 8,879 7,563 8,381 8,349 8,195 8,423 9,191 7,757 8,636 8,603 8,444 8,679 9,515 7,957 8,808 8,775 8,613 8,853 9,752 8,077 8,984 8,950 8,785 9,030 9,996 8,198 9,164 9,129 8,960 9,210 10,246 8,321 9,347 9,312 9,140 9,394 10,502 8,445 9,534 9,498 9,322 9,582 10,765 8,572 9,725 9,688 9,509 9,774 11,034 8,701 9,920 9,882 9,699 9,969 11,310 8,831 10,118 10,080 9,893 10,169 11,593 8,964 Steer Davies Gleave Page 245 of 259

146 Traffic study for project Conexión Antioquia - Bolívar Year Base scenario SI SII SIII SIV SV ,320 10,281 10,091 10,372 11,882 9,098 10,527 10,487 10,293 10,580 12,179 9,235 10,737 10,697 10,499 10,791 12,484 9,373 10,952 10,911 10,709 11,007 12,796 9,514 11,171 11,129 10,923 11,227 13,116 9,656 11,395 11,351 11,141 11,452 13,444 9,801 11,622 11,578 11,364 11,681 13,780 9,948 11,855 11,810 11,591 11,914 14,124 10,098 12,092 12,046 11,823 12,153 14,478 10,249 12,334 12,287 12,060 12,396 14,840 10,403 12,580 12,533 12,301 12,644 15,211 10,559 12,832 12,783 12,547 12,897 15,591 10,717 Note: Traffic projections of this station might vary according to the ultimate location of the toll plaza. These projections consider the location of Purgatorio according to what is stated in the concession contract for the complete calendar year of Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Table 11-38: Percentage change between sensitivities and base case Purgatorio Year SI SII SIII SIV SV % 0% 0% 2% -1% % 0% 0% 2% -1% % 0% 0% 3% -2% % 0% 0% 3% -2% % 0% 0% 4% -2% % 0% 0% 4% -3% % 0% 0% 5% -3% % -2% 0% 5% -4% % -2% 0% 6% -4% % -2% 0% 6% -5% % -2% 0% 7% -5% % -2% 0% 7% -5% % -2% 0% 8% -6% % -2% 0% 8% -6% % -2% 0% 9% -7% Steer Davies Gleave Page 246 of 259

147 Traffic study for project Conexión Antioquia - Bolívar Year SI SII SIII SIV SV % -2% 0% 9% -7% % -2% 1% 10% -7% % -2% 0% 10% -8% % -2% 1% 11% -8% % -2% 1% 11% -9% % -2% 1% 12% -9% % -2% 1% 12% -10% % -2% 1% 13% -10% % -2% 1% 13% -11% % -2% 0% 14% -11% % -2% 1% 15% -11% % -2% 1% 15% -12% % -2% 1% 16% -12% % -2% 1% 16% -13% % -2% 1% 17% -13% % -2% 1% 17% -14% % -2% 1% 18% -14% % -2% 1% 19% -14% % -2% 0% 19% -15% % -2% 1% 20% -15% % -2% 1% 20% -16% % -2% 1% 21% -16% % -2% 1% 22% -16% Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 247 of 259

148 Traffic study for project Conexión Antioquia - Bolívar Table 11-39: Traffic projections, sensitivity tests San Carlos Year Base scenario SI SII SIII SIV SV , ,152 1,045 1, ,197 1,073 1, ,244 1,102 1, ,293 1,132 1, ,343 1,163 1, ,396 1,195 1, ,446 1,227 1, ,498 1,261 1, ,551 1,296 1, ,607 1,332 1, ,665 1,369 1, ,706 1,389 1, ,749 1,410 1, ,002 1,793 1,431 1, ,022 1,838 1,453 1, ,043 1,884 1,474 1,701 1, ,063 1,931 1,496 1,735 1, ,085 1,979 1,519 Steer Davies Gleave Page 248 of 259

149 Traffic study for project Conexión Antioquia - Bolívar Year Base scenario SI SII SIII SIV SV 1,770 1, ,106 2,028 1, ,805 1, ,128 2,079 1,565 1,841 1, ,151 2,131 1,588 1,878 1, ,174 2,184 1,612 1,916 1, ,198 2,239 1,636 1,954 1, ,221 2,295 1,661 1,993 1, ,246 2,352 1,686 2,033 1, ,271 2,411 1,711 2,074 1,221 1,006 1,296 2,471 1,737 2,115 1,245 1,026 1,322 2,533 1,763 2,158 1,270 1,047 1,349 2,596 1,789 2,201 1,295 1,068 1,376 2,661 1,816 2,245 1,321 1,089 1,403 2,728 1,843 Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Table 11-40: Percentage change between sensitivities and base case San Carlos Year SI SII SIII SIV SV % 0% 0% 0% 0% % 0% 0% 2% -2% % 0% 0% 2% -2% % -10% 0% 3% -3% % -10% 0% 4% -4% % -9% 0% 5% -4% % -9% 0% 5% -5% % -54% -40% 5% -5% % -54% -40% 5% -5% % -53% -39% 6% -6% % -53% -39% 7% -6% % -53% -39% 7% -7% % -52% -38% 8% -7% % -52% -38% 9% -8% % -52% -38% 9% -8% % -52% -38% 9% -9% % -52% -38% 10% -9% Steer Davies Gleave Page 249 of 259

150 Traffic study for project Conexión Antioquia - Bolívar Year SI SII SIII SIV SV % -51% -38% 10% -9% % -51% -38% 11% -10% % -51% -38% 11% -10% % -51% -37% 12% -11% % -51% -37% 12% -11% % -51% -37% 13% -12% % -51% -38% 14% -12% % -51% -37% 14% -12% % -51% -38% 15% -13% % -51% -38% 15% -13% % -51% -37% 16% -14% % -51% -37% 16% -14% % -51% -37% 17% -15% % -51% -38% 17% -15% % -51% -37% 18% -15% % -52% -37% 19% -16% % -51% -38% 19% -16% % -51% -37% 20% -17% % -51% -37% 20% -17% % -51% -37% 21% -17% % -51% -38% 22% -18% Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 250 of 259

151 Traffic study for project Conexión Antioquia - Bolívar Tabla 11-41: Traffic projections, sensitivity tests Mata de Caña Year Base scenario SI SII SIII SIV SV ,815 2,815 2,815 2,815 2,858 2,780 2,929 2,929 2,929 2,929 2,989 2,883 3,060 3,060 3,060 3,060 3,142 3,001 3,693 3,664 3,664 3,693 3,811 3,602 3,828 3,801 3,801 3,828 3,970 3,716 3,947 3,914 3,914 3,947 4,114 3,814 4,094 4,058 4,058 4,094 4,286 3,937 4,887 4,224 4,091 4,269 5,134 4,680 5,043 4,365 4,226 4,410 5,325 4,806 5,204 4,511 4,366 4,557 5,522 4,936 5,370 4,661 4,510 4,708 5,728 5,069 5,542 4,817 4,659 4,864 5,941 5,206 5,719 4,979 4,814 5,025 6,162 5,347 5,897 5,142 4,969 5,189 6,385 5,488 6,081 5,310 5,130 5,357 6,616 5,632 6,271 5,484 5,296 5,531 6,855 5,780 6,467 5,664 5,467 5,711 7,104 5,932 6,669 5,850 5,644 5,897 7,361 6,089 6,802 5,967 5,757 6,015 7,545 6,180 6,938 6,086 5,873 6,135 7,734 6,273 7,077 6,208 5,990 6,258 7,927 6,367 7,218 6,332 6,110 6,383 8,125 6,462 7,363 6,459 6,232 6,511 8,328 6,559 7,510 6,588 6,357 6,641 8,537 6,657 7,660 6,720 6,484 6,774 8,750 6,757 7,813 6,854 6,613 6,909 8,969 6,859 Steer Davies Gleave Page 251 of 259

152 Traffic study for project Conexión Antioquia - Bolívar Year Base scenario SI SII SIII SIV SV ,970 6,991 6,746 7,047 9,193 6,962 8,129 7,131 6,881 7,188 9,423 7,066 8,292 7,274 7,018 7,332 9,658 7,172 8,457 7,419 7,159 7,479 9,900 7,280 8,626 7,567 7,302 7,628 10,147 7,389 8,799 7,719 7,448 7,781 10,401 7,500 8,975 7,873 7,597 7,936 10,661 7,612 9,155 8,031 7,749 8,095 10,928 7,726 9,338 8,191 7,904 8,257 11,201 7,842 9,524 8,355 8,062 8,422 11,481 7,960 9,715 8,522 8,223 8,590 11,768 8, ,909 8,693 8,387 8,762 12,062 8,200 Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Table 11-42: Percentage change between sensitivities and base case Mata de Caña Year SI SII SIII SIV SV % 0% 0% 2% -1% % 0% 0% 2% -2% % 0% 0% 3% -2% % -1% 0% 3% -2% % -1% 0% 4% -3% % -1% 0% 4% -3% % -1% 0% 5% -4% % -16% -13% 5% -4% % -16% -13% 6% -5% % -16% -12% 6% -5% % -16% -12% 7% -6% % -16% -12% 7% -6% % -16% -12% 8% -7% % -16% -12% 8% -7% % -16% -12% 9% -7% % -16% -12% 9% -8% % -15% -12% 10% -8% % -15% -12% 10% -9% Steer Davies Gleave Page 252 of 259

153 Traffic study for project Conexión Antioquia - Bolívar Year SI SII SIII SIV SV % -15% -12% 11% -9% % -15% -12% 11% -10% % -15% -12% 12% -10% % -15% -12% 13% -10% % -15% -12% 13% -11% % -15% -12% 14% -11% % -15% -12% 14% -12% % -15% -12% 15% -12% % -15% -12% 15% -13% % -15% -12% 16% -13% % -15% -12% 16% -14% % -15% -12% 17% -14% % -15% -12% 18% -14% % -15% -12% 18% -15% % -15% -12% 19% -15% % -15% -12% 19% -16% % -15% -12% 20% -16% % -15% -12% 21% -16% % -15% -12% 21% -17% % -15% -12% 22% -17% Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 253 of 259

154 Traffic study for project Conexión Antioquia - Bolívar Table 11-43: Traffic projections, sensitivity tests Caimanera Year Base scenario SI SII SIII SIV SV ,433 2,433 2,433 2,433 2,473 2,405 2,511 2,511 2,511 2,511 2,567 2,474 2,615 2,615 2,615 2,615 2,692 2,569 3,402 3,353 3,353 3,402 3,521 3,321 3,556 3,508 3,508 3,556 3,699 3,456 3,667 3,606 3,606 3,667 3,836 3,546 3,851 3,785 3,785 3,851 4,044 3,707 4,620 3,888 3,672 3,965 4,868 4,424 4,762 4,014 3,789 4,092 5,044 4,537 4,909 4,145 3,909 4,224 5,228 4,654 5,061 4,280 4,034 4,360 5,419 4,773 5,217 4,420 4,163 4,500 5,618 4,896 5,379 4,565 4,297 4,646 5,824 5,023 5,543 4,712 4,431 4,794 6,032 5,150 5,713 4,864 4,571 4,946 6,248 5,281 5,887 5,021 4,715 5,104 6,472 5,415 6,068 5,184 4,864 5,268 6,705 5,553 6,254 5,352 5,018 5,436 6,947 5,694 6,374 5,454 5,113 5,540 7,114 5,776 6,497 5,558 5,211 5,646 7,286 5,859 6,622 5,665 5,310 5,754 7,462 5,943 6,749 5,773 5,411 5,864 7,642 6,028 6,879 5,883 5,514 5,976 7,827 6,115 7,012 5,996 5,620 6,091 8,017 6,203 7,147 6,111 5,727 6,208 8,211 6,292 Steer Davies Gleave Page 254 of 259

155 Traffic study for project Conexión Antioquia - Bolívar Year Base scenario SI SII SIII SIV SV 7,285 6,228 5,837 6,327 8,410 6, ,426 6,348 5,949 6,449 8,614 6,475 7,569 6,470 6,063 6,573 8,823 6,568 7,716 6,594 6,179 6,699 9,038 6,663 7,865 6,721 6,298 6,828 9,258 6,759 8,018 6,851 6,419 6,960 9,483 6,857 8,173 6,983 6,542 7,094 9,714 6,956 8,332 7,118 6,668 7,231 9,951 7,057 8,493 7,255 6,796 7,371 10,193 7,159 8,658 7,395 6,927 7,513 10,442 7,262 8,827 7,538 7,061 7,659 10,697 7,367 8,998 7,684 7,197 7,807 10,958 7,474 9,173 7,833 7,336 7,958 11,226 7,583 Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Table 11-44: Percentage change between sensitivities and base case Caimanera Year SI SII SIII SIV SV % 0% 0% 2% -1% % 0% 0% 2% -1% % 0% 0% 3% -2% % -1% 0% 3% -2% % -1% 0% 4% -3% % -2% 0% 5% -3% % -2% 0% 5% -4% % -21% -14% 5% -4% % -20% -14% 6% -5% % -20% -14% 6% -5% % -20% -14% 7% -6% % -20% -14% 8% -6% % -20% -14% 8% -7% % -20% -14% 9% -7% % -20% -13% 9% -8% % -20% -13% 10% -8% Steer Davies Gleave Page 255 of 259

156 Traffic study for project Conexión Antioquia - Bolívar Year SI SII SIII SIV SV % -20% -13% 10% -8% % -20% -13% 11% -9% % -20% -13% 12% -9% % -20% -13% 12% -10% % -20% -13% 13% -10% % -20% -13% 13% -11% % -20% -13% 14% -11% % -20% -13% 14% -12% % -20% -13% 15% -12% % -20% -13% 15% -12% % -20% -13% 16% -13% % -20% -13% 17% -13% % -20% -13% 17% -14% % -20% -13% 18% -14% % -20% -13% 18% -14% % -20% -13% 19% -15% % -20% -13% 19% -15% % -20% -13% 20% -16% % -20% -13% 21% -16% % -20% -13% 21% -17% % -20% -13% 22% -17% % -20% -13% 22% -17% Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 256 of 259

157 Traffic study for project Conexión Antioquia - Bolívar Table 11-45: Traffic projections, sensitivity tests San Onofre Year Base scenario SI SII SIII SIV SV ,125 2,125 2,125 2,125 2,157 2,097 2,114 2,114 2,114 2,114 2,157 2,078 2,172 2,172 2,172 2,172 2,230 2,126 2,783 2,780 2,780 2,783 2,870 2,710 2,796 2,799 2,799 2,796 2,899 2,709 2,993 2,989 2,989 2,993 3,120 2,887 3,012 3,009 3,009 3,012 3,153 2,888 3,276 3,218 2,905 3,230 3,441 3,130 3,381 3,324 3,000 3,337 3,570 3,215 3,490 3,435 3,098 3,448 3,704 3,302 3,602 3,549 3,200 3,562 3,843 3,392 3,717 3,667 3,304 3,680 3,987 3,484 3,837 3,789 3,412 3,803 4,137 3,578 3,958 3,913 3,522 3,926 4,291 3,673 4,083 4,040 3,635 4,054 4,451 3,770 4,213 4,172 3,752 4,187 4,617 3,870 4,346 4,308 3,872 4,323 4,789 3,973 4,484 4,449 3,997 4,464 4,969 4,078 4,573 4,538 4,077 4,553 5,093 4,139 4,665 4,629 4,158 4,644 5,220 4,201 4,758 4,721 4,241 4,737 5,351 4,264 4,853 4,816 4,326 4,832 5,484 4,328 4,950 4,912 4,413 4,929 5,621 4,393 5,049 5,010 4,501 5,027 5,762 4,459 5,150 5,111 4,591 5,128 5,906 4,526 5,253 5,213 4,683 5,230 6,054 4,594 Steer Davies Gleave Page 257 of 259

158 Traffic study for project Conexión Antioquia - Bolívar Year Base scenario SI SII SIII SIV SV ,358 5,317 4,776 5,335 6,205 4,663 5,465 5,423 4,872 5,442 6,360 4,733 5,575 5,532 4,969 5,550 6,519 4,804 5,686 5,643 5,069 5,661 6,682 4,876 5,800 5,755 5,170 5,775 6,849 4,949 5,916 5,871 5,274 5,890 7,020 5,023 6,034 5,988 5,379 6,008 7,196 5,098 6,155 6,108 5,487 6,128 7,376 5,175 6,278 6,230 5,596 6,251 7,560 5,253 6,404 6,354 5,708 6,376 7,749 5,331 6,532 6,482 5,823 6,503 7,943 5,411 6,662 6,611 5,939 6,633 8,141 5,492 Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Table 11-46: Percentage change between sensitivities and base case San Onofre Year SI SII SIII SIV SV % 0% 0% 2% -1% % 0% 0% 2% -2% % 0% 0% 3% -2% % 0% 0% 3% -3% % 0% 0% 4% -3% % 0% 0% 4% -4% % 0% 0% 5% -4% % -11% -1% 5% -4% % -11% -1% 6% -5% % -11% -1% 6% -5% % -11% -1% 7% -6% % -11% -1% 7% -6% % -11% -1% 8% -7% % -11% -1% 8% -7% % -11% -1% 9% -8% % -11% -1% 10% -8% % -11% -1% 10% -9% % -11% 0% 11% -9% Steer Davies Gleave Page 258 of 259

159 Traffic study for project Conexión Antioquia - Bolívar Year SI SII SIII SIV SV % -11% 0% 11% -9% % -11% 0% 12% -10% % -11% 0% 12% -10% % -11% 0% 13% -11% % -11% 0% 14% -11% % -11% 0% 14% -12% % -11% 0% 15% -12% % -11% 0% 15% -13% % -11% 0% 16% -13% % -11% 0% 16% -13% % -11% 0% 17% -14% % -11% 0% 18% -14% % -11% 0% 18% -15% % -11% 0% 19% -15% % -11% 0% 19% -16% % -11% 0% 20% -16% % -11% 0% 20% -16% % -11% 0% 21% -17% % -11% 0% 22% -17% % -11% 0% 22% -18% Source: Steer Davies Gleave, forecasts March 2017, traffic does not differentiate for discounted tariffs or by vehicles bypassing the toll plazas. Steer Davies Gleave Page 259 of 259

160 [THIS PAGE INTENTIONALLY LEFT BLANK]

161 B APPENDIX B INDEPENDENT ENGINEERS REPORT B-1

162 [THIS PAGE INTENTIONALLY LEFT BLANK]

163

164 Arup Latin America S.A.. Final IE Report December 13, 2017

165 Arup Latin America S.A.. Final IE Report December 13, 2017

166 Arup Latin America S.A.. Final IE Report December 13, 2017

167 Arup Latin America S.A.. Final IE Report December 13, 2017

168 [THIS PAGE INTENTIONALLY LEFT BLANK]

169 Arup Latin America S.A. Final IE Report December 13, 2017 Page 1

170 Arup Latin America S.A. Final IE Report December 13, 2017 Page 2

171 Arup Latin America S.A. Final IE Report December 13, 2017 Page 3

172 Arup Latin America S.A. Final IE Report December 13, 2017 Page 4

173 Arup Latin America S.A. Final IE Report December 13, 2017 Page 5

174 Arup Latin America S.A. Final IE Report December 13, 2017 Page 6

175 Arup Latin America S.A. Final IE Report December 13, 2017 Page 7

176 Arup Latin America S.A. Final IE Report December 13, 2017 Page 8

177 Arup Latin America S.A. Final IE Report December 13, 2017 Page 9

178 Arup Latin America S.A. Final IE Report December 13, 2017 Page 10

179 Arup Latin America S.A. Final IE Report December 13, 2017 Page 11

180 Arup Latin America S.A. Final IE Report December 13, 2017 Page 12

181 Arup Latin America S.A. Final IE Report December 13, 2017 Page 13

182 Arup Latin America S.A. Final IE Report December 13, 2017 Page 14

183 Arup Latin America S.A. Final IE Report December 13, 2017 Page 15

184 Arup Latin America S.A. Final IE Report December 13, 2017 Page 16

185 Arup Latin America S.A. Final IE Report December 13, 2017 Page 17

186 Arup Latin America S.A. Final IE Report December 13, 2017 Page 18

187 Arup Latin America S.A. Final IE Report December 13, 2017 Page 19

188 Arup Latin America S.A. Final IE Report December 13, 2017 Page 20

189 Arup Latin America S.A. Final IE Report December 13, 2017 Page 21

190 Arup Latin America S.A. Final IE Report December 13, 2017 Page 22

191 Arup Latin America S.A. Final IE Report December 13, 2017 Page 23

192 Arup Latin America S.A. Final IE Report December 13, 2017 Page 24

193 Arup Latin America S.A. Final IE Report December 13, 2017 Page 25

194 Arup Latin America S.A. Final IE Report December 13, 2017 Page 26

195 Arup Latin America S.A. Final IE Report December 13, 2017 Page 27

196 Arup Latin America S.A. Final IE Report December 13, 2017 Page 28

197 Arup Latin America S.A. Final IE Report December 13, 2017 Page 29

198 Arup Latin America S.A. Final IE Report December 13, 2017 Page 30

199 Arup Latin America S.A. Final IE Report December 13, 2017 Page 31

200 Arup Latin America S.A. Final IE Report December 13, 2017 Page 32

201 Arup Latin America S.A. Final IE Report December 13, 2017 Page 33

202 Arup Latin America S.A. Final IE Report December 13, 2017 Page 34

203 Arup Latin America S.A. Final IE Report December 13, 2017 Page 35

204 Arup Latin America S.A. Final IE Report December 13, 2017 Page 36

205 Arup Latin America S.A. Final IE Report December 13, 2017 Page 37

206 Arup Latin America S.A. Final IE Report December 13, 2017 Page 38

207 Arup Latin America S.A. Final IE Report December 13, 2017 Page 39

208 Arup Latin America S.A. Final IE Report December 13, 2017 Page 40

209 Arup Latin America S.A. Final IE Report December 13, 2017 Page 41

210 Arup Latin America S.A. Final IE Report December 13, 2017 Page 42

211 Arup Latin America S.A. Final IE Report December 13, 2017 Page 43

212 Arup Latin America S.A. Final IE Report December 13, 2017 Page 44

213 Arup Latin America S.A. Final IE Report December 13, 2017 Page 45

214 Arup Latin America S.A. Final IE Report December 13, 2017 Page 46

215 Arup Latin America S.A. Final IE Report December 13, 2017 Page 47

216 Arup Latin America S.A. Final IE Report December 13, 2017 Page 48

217 Arup Latin America S.A. Final IE Report December 13, 2017 Page 49

218 Arup Latin America S.A. Final IE Report December 13, 2017 Page 50

219 Arup Latin America S.A. Final IE Report December 13, 2017 Page 51

220 Arup Latin America S.A. Final IE Report December 13, 2017 Page 52

221 Arup Latin America S.A. Final IE Report December 13, 2017 Page 53

222 Arup Latin America S.A. Final IE Report December 13, 2017 Page 54

223 o o o Arup Latin America S.A. Final IE Report December 13, 2017 Page 55

224 Arup Latin America S.A. Final IE Report December 13, 2017 Page 56

225 Arup Latin America S.A. Final IE Report December 13, 2017 Page 57

226 Arup Latin America S.A. Final IE Report December 13, 2017 Page 58

227 Arup Latin America S.A. Final IE Report December 13, 2017 Page 59

228 Arup Latin America S.A. Final IE Report December 13, 2017 Page 60

229 Arup Latin America S.A. Final IE Report December 13, 2017 Page 61

230 Arup Latin America S.A. Final IE Report December 13, 2017 Page 62

231 Arup Latin America S.A. Final IE Report December 13, 2017 Page 63

232 Arup Latin America S.A. Final IE Report December 13, 2017 Page 64

233 Arup Latin America S.A. Final IE Report December 13, 2017 Page 65

234 Arup Latin America S.A. Final IE Report December 13, 2017 Page 66

235 Arup Latin America S.A. Final IE Report December 13, 2017 Page 67

236 Arup Latin America S.A. Final IE Report December 13, 2017 Page 68

237 Arup Latin America S.A. Final IE Report December 13, 2017 Page 69

238 Arup Latin America S.A. Final IE Report December 13, 2017 Page 70

239 Arup Latin America S.A. Final IE Report December 13, 2017 Page 71

240 Arup Latin America S.A. Final IE Report December 13, 2017 Page 72

241 Arup Latin America S.A. Final IE Report December 13, 2017 Page 73

242 Arup Latin America S.A. Final IE Report December 13, 2017 Page 74

243 Arup Latin America S.A. Final IE Report December 13, 2017 Page 75

244 Arup Latin America S.A. Final IE Report December 13, 2017 Page 76

245 Arup Latin America S.A. Final IE Report December 13, 2017 Page 77

246 Arup Latin America S.A. Final IE Report December 13, 2017 Page 78

247 Arup Latin America S.A. Final IE Report December 13, 2017 Page 79

248 Arup Latin America S.A. Final IE Report December 13, 2017 Page 80

249 Arup Latin America S.A. Final IE Report December 13, 2017 Page 81

250 Arup Latin America S.A. Final IE Report December 13, 2017 Page 82

251 Arup Latin America S.A. Final IE Report December 13, 2017 Page 83

252 Arup Latin America S.A. Final IE Report December 13, 2017 Page 84

253 Arup Latin America S.A. Final IE Report December 13, 2017 Page 85

254 Arup Latin America S.A. Final IE Report December 13, 2017 Page 86

255 Arup Latin America S.A. Final IE Report December 13, 2017 Page 87

256 Arup Latin America S.A. Final IE Report December 13, 2017 Page 88

257 Arup Latin America S.A. Final IE Report December 13, 2017 Page 89

258 Arup Latin America S.A. Final IE Report December 13, 2017 Page 90

259 Arup Latin America S.A. Final IE Report December 13, 2017 Page 91

260 Arup Latin America S.A. Final IE Report December 13, 2017 Page 92

261 Arup Latin America S.A. Final IE Report December 13, 2017 Page 93

262 Arup Latin America S.A. Final IE Report December 13, 2017 Page 94

263 Arup Latin America S.A. Final IE Report December 13, 2017 Page 95

264 Arup Latin America S.A. Final IE Report December 13, 2017 Page 96

265 Arup Latin America S.A. Final IE Report December 13, 2017 Page 97

266 Arup Latin America S.A. Final IE Report December 13, 2017 Page 98

267 Arup Latin America S.A. Final IE Report December 13, 2017 Page 99

268 Arup Latin America S.A. Final IE Report December 13, 2017 Page 100

269 Arup Latin America S.A. Final IE Report December 13, 2017 Page 101

270 Arup Latin America S.A. Final IE Report December 13, 2017 Page 102

271 Arup Latin America S.A. Final IE Report December 13, 2017 Page 103

272 Arup Latin America S.A. Final IE Report December 13, 2017 Page 104

273 Arup Latin America S.A. Final IE Report December 13, 2017 Page 105

274 Arup Latin America S.A. Final IE Report December 13, 2017 Page 106

275 Arup Latin America S.A. Final IE Report December 13, 2017 Page 107

276 Arup Latin America S.A. Final IE Report December 13, 2017 Page 108

277 Arup Latin America S.A. Final IE Report December 13, 2017 Page 109

278 Arup Latin America S.A. Final IE Report December 13, 2017 Page 110

279 Arup Latin America S.A. Final IE Report December 13, 2017 Page 111

280 Poliza Arup Latin America S.A. Final IE Report December 13, 2017 Page 112

281 Arup Latin America S.A. Final IE Report December 13, 2017 Page 113

282 Arup Latin America S.A. Final IE Report December 13, 2017 Page 114

283 Arup Latin America S.A. Final IE Report December 13, 2017 Page 115

284 Arup Latin America S.A. Final IE Report December 13, 2017 Page 116

285 Arup Latin America S.A. Final IE Report December 13, 2017 Page 117

286 Arup Latin America S.A. Final IE Report December 13, 2017 Page 118

287 Arup Latin America S.A. Final IE Report December 13, 2017 Page 119

288 Arup Latin America S.A. Final IE Report December 13, 2017 Page 120

289 Arup Latin America S.A. Final IE Report December 13, 2017 Page 121

290 Arup Latin America S.A. Final IE Report December 13, 2017 Page 122

291 Arup Latin America S.A. Final IE Report December 13, 2017 Page 123

292 Arup Latin America S.A. Final IE Report December 13, 2017 Page 124

293 Arup Latin America S.A. Final IE Report December 13, 2017 Page 125

294 Arup Latin America S.A. Final IE Report December 13, 2017 Page 126

295 Arup Latin America S.A. Final IE Report December 13, 2017 Page 127

296 Arup Latin America S.A. Final IE Report December 13, 2017 Page 128

297 Arup Latin America S.A. Final IE Report December 13, 2017 Page 129

298 Arup Latin America S.A. Final IE Report December 13, 2017 Page 130

299 Arup Latin America S.A. Final IE Report December 13, 2017 Page 131

300 Arup Latin America S.A. Final IE Report December 13, 2017 Page 132

301 Arup Latin America S.A. Final IE Report December 13, 2017 Page 133

302 Arup Latin America S.A. Final IE Report December 13, 2017 Page 134

303 Arup Latin America S.A. Final IE Report December 13, 2017 Page 135

304 Arup Latin America S.A. Final IE Report December 13, 2017 Page 136

305 Arup Latin America S.A. Final IE Report December 13, 2017 Page 137

306 Arup Latin America S.A. Final IE Report December 13, 2017 Page 138

307 Arup Latin America S.A. Final IE Report December 13, 2017 Page 139

308 Arup Latin America S.A. Final IE Report December 13, 2017 Page 140

309 Arup Latin America S.A. Final IE Report December 13, 2017 Page 141

310 Arup Latin America S.A. Final IE Report December 13, 2017 Page 142

311 Arup Latin America S.A. Final IE Report December 13, 2017 Page 143

312 Arup Latin America S.A. Final IE Report December 13, 2017 Page 144

313 Arup Latin America S.A. Final IE Report December 13, 2017 Page 145

314 Arup Latin America S.A. Final IE Report December 13, 2017 Page 146

315 Arup Latin America S.A. Final IE Report December 13, 2017 Page 147

316 Arup Latin America S.A. Final IE Report December 13, 2017 Page 148

317 Arup Latin America S.A. Final IE Report December 13, 2017 Page 149

318 Arup Latin America S.A. Final IE Report December 13, 2017 Page 150

319 Arup Latin America S.A. Final IE Report December 13, 2017 Page 151

320 Arup Latin America S.A. Final IE Report December 13, 2017 Page 152

321 Arup Latin America S.A. Final IE Report December 13, 2017 Page 153

322 Arup Latin America S.A. Final IE Report December 13, 2017 Page 154

323 Arup Latin America S.A. Final IE Report December 13, 2017 Page 155

324 Arup Latin America S.A. Final IE Report December 13, 2017 Page 156

325 Arup Latin America S.A. Final IE Report December 13, 2017 Page 157

326 Arup Latin America S.A. Final IE Report December 13, 2017 Page 158

327 Arup Latin America S.A. Final IE Report December 13, 2017 Page 159

328 Arup Latin America S.A. Final IE Report December 13, 2017 Page 160

329 Arup Latin America S.A. Final IE Report December 13, 2017 Page 161