Documento sin título

Superpave Zoning for Chile


R. Delgadillo *, M. Segovia *, C. Wahr 1*, G. Thenoux **


* Universidad Técnica Federico Santa María, Valparaíso. CHILE

** Pontificia Universidad Católica de Chile, Santiago. CHILE

Corresponding author


In Chile, the selection of asphalt binders is based on traditional specifications. Although the Superpave specification has not been implemented yet, it was consider important to make a zoning process of the Chilean territory according to this methodology. This activity relied on the information of 94 weather stations of the Chilean Meteorological Service (DMC in Spanish) and Chile’s Water Department (DGA in Spanish), which have reliable data for a minimum of 20 years. Weather data, together with the Köppen climate classification for Chile and the topography of our territory, were used to define approximate zones where the use of each type of asphalt binder is appropriate. Zoning data indicate that most of our territory can be covered by three types of asphalt binders: for the northern and southern regions the use of PG 58-28 is recommended; the central region requires PG 64-22; and finally, the Patagonia and high mountain zones need PG 52-34. In the IX Region, there is a small area in the Andean foothills, where PG 64-34 is required, according to the available climate information and the methodology applied. Classifications PG 64-22, PG 58-28 and PG 52-34 are traditional binders, which were zoned for high-speed traffic conditions and moderate traffic volume.

Keywords: Pavement maintenance, superpave, structural layers, asphalt binder, influence of weather station

1. Introduction

Pavements’ maintenance requirements greatly depend on the adequate selection of materials for their respective structural layers. Asphalt binders are susceptible to thermal conditions, that is, their performance strongly depends on the existing weather. For example, a binder can have enough stiffness to resist rutting in a cold zone, but it can show a bad performance in a warm zone concerning the same failure. In recent years, relevant researches have been made in Chile, which address the influence of climate and the advanced characterization of asphalt materials on the performance and maintenance of flexible pavements (Delgadillo et al., 2011; Araya et al., 2012; Delgadillo et al., 2014; García et al., 2014; Osorio et al., 2015).

Considering the above, in 1987, the Strategic Highway Research Program, SHRP, (Kennedy et al., 1994) of the United States began to develop a new specification system for asphalt materials. The final product of the research project was a system known as Superpave (SUperior PERforming Asphalt PAVEments). The main differences and advantages of the new specification in relation to the traditional one are the following:

The material is specified considering the temperatures expected on site.

• Basic properties of the binder are measured by instruments (rheometers).

• It considers the long-term aging of the material.

• It considers the load time effect (indirectly).

• It considers the traffic volume effect (indirectly).

Binders are characterized according to the performance expected by three types of failure: rutting, fatigue and thermal cracking. Since each of these failures occur at different temperatures, the asphalt denomination has three characteristic temperatures, expressed in Celsius degrees. For example, a binder labeled PG 64-22 has the following properties:

• It resists rutting at temperatures up to Txx = 64°C.

• It resists thermal rutting at temperatures up to Tyy = -22°C.

• It resists fatigue at temperatures less or equal to


A detailed description of the specifications, including the equipment to be used and parameters to be controlled, can be found in different references (AASHTO, 2015; Asphalt Institute, 2003; ASTM 2015).

The use of Superpave specifications requires to know the pavement temperatures expected on the site’s region. This allows selecting a binder that is adequate for that climate zone. This is especially important in a country with such climate diversity as Chile. Previous Superpave zoning efforts for binders were made by Vivanco and Bahía (2005) and by Contreras (2007). The first work used general and very limited climate information, without a statistical analysis of extreme temperatures from the weather stations, which turns it into a more theoretical exercise rather than a practical result. The second work determined PG temperatures of the area between Santiago and Los Angeles, based on data from 38 weather stations, but no specific recommendations were made regarding the use of traditional binders in each zone.

The present work carries out a Superpave zoning in the entire Chilean continental territory, based on the original methodology of the SHRP project (Huber 1994). The outcome of this work is a territorial division by climate zones, which assigns the corresponding traditional asphalt binder to each one of them.

2. Weather station selection

The available information included a total of 137 weather stations with data of extreme daily temperatures, 106 coming from Chile’s Water Department (DGA, ) and 31 from the Chilean Meteorological Service (DMC, ). However, a minimum of 20 consecutive years of reliable data were needed for the information to be representative and with statistic validity for determining reliabilities. Therefore, it was necessary to apply certain minimum selection criteria, which are detailed below.

Criterion 1: A Complete Year of Relevant Data

The preliminary selection of the PG grade was made based on extreme temperatures; consequently, the stations to be selected should have quality information of the minimum and maximum temperatures. Chile is a country with annual temperature cycles, so it can be assumed that maximum daily temperatures will occur in the summer season, while minimum ones will be registered during the winter months. Considering this, the first two filters defined to select a station were the following:

• Daily records of minimum temperature for 95% of the days comprised between May 21 and September 21.

• Daily records of maximum temperature for 95% of the days comprised between November 21 and March 21.

Criterion 2: Pending Days for the Maximum Temperature

The relevant temperature for rutting is the maximum moving average of seven consecutive days. Therefore, it was necessary to define an additional criterion for those days with no information on maximum temperature, which consisted in eliminating all those years having two or more days without maximum temperature within any given period of seven consecutive days. In relation to the years having only one day with no record within a given period of seven consecutive days, its value was interpolated between the previous and following day.

Criterion 3: Updated Information

Considering that climate has suffered significant changes in the last 100 years, it was judged that updated data was also a relevant criteria for selecting the stations. Thus, the final selection only included stations that, besides meeting criteria 1 and 2, had information until at least the year 2005.

Total Selected Stations

The total number of weather stations that complied with the aforementioned criteria was 94. This meant to dismiss 43 of the 137 stations originally considered, but this ensured a greater data reliability.

3. PG grade calculation of the selected stations

The extreme temperatures expected on the pavement were calculated, which defines the applicable PG grade for each selected weather station. The original formulas of the Superpave methodology were used (Huber 1994(, which generally yield more conservative results than those of the Long Term Pavement Performance formulas (Mohseni 1998).

Regarding the pavement’s minimum temperature, it is assumed that it is the same as the minimum air temperature, so the steps followed by each station were:

1) Select the minimum (air) temperature daily record for the available year.

2) Average the yearly minimum temperatures selected in step 1, which defines the minimum 50% temperature reliability.

3) Calculate the deviation of the records selected in step 1, which allows defining the minimum temperatures for other reliabilities.

The pavement’s maximum temperature is calculated based on the maximum air temperature using the following formula (Huber 1994):


Where Tair is the maximum air temperature, Tarea is the maximum temperature of the pavement surface. Then, the temperature is calculated at 20 mm depth from the surface T20 mm, which is relevant to define the PG grade, as follows:


Where φ is the absolute value of the station’s latitude in degrees. The following steps to calculate the each station’s maximum temperatures are:

4. Calculate the average maximum (air) temperature of seven consecutive days for each year.

5. Average the temperatures calculated in step 4, which defines the maximum air temperature Tair at 50% reliability.

6. Calculate the deviation of temperatures calculated in step 4, which allows defining the maximum temperatures for other reliabilities.

7. Calculate the maximum daily temperature of the pavement surface Tsurface, entering the Tair calculated in step 5 in equation 1.

8. Calculate the maximum temperature at 20 [mm] depth, entering the Tsurface calculated in step 7 in equation 2.

9. Calculate the standard deviation of the maximum temperature at 20 mm depth from the surface, using the standard deviation of the air temperature calculated in step 6, plus formulas 1 and 2.

The PG XX-YY grade at 50% reliability for each station will be given by the temperatures calculated in steps 2 (YY) and 8 (XX). If a higher reliability is desired, these temperatures need to be modified using the corresponding standard deviation. For example, at 98% reliability, Superpave recommends subtracting 2 times the standard deviation calculated in step 3 from the temperature YY and add 2 times the standard deviation calculated in step 9 to the temperature XX.

The Superpave classification considers a discrete scale with increases every 6°C; therefore, the obtained temperatures must be approximated to the available grade immediately higher or lower, according to the maximum and minimum temperatures, respectively. Columns Txx and Tyy of Table 1 show the PG grades for each station at 50% and 98% reliabilities.

The extreme temperatures Txx and Tyy determine the requirements to be met by the binder, with the aim of reducing the rutting and thermal cracking susceptibility. Both extreme temperatures are associated to a medium temperature that is characteristic of the Tint location, which is calculated according to the following equation:


The Tint columns of Table 1 show the temperatures of each station at 50% and 98% reliabilities.

4. Simplified zoning using traditional binders

Selection of Appropriate Binder for each Zone

The previous results show the minimum extreme and medium temperature requirements that binders used in each locality have to meet. Given the large climate variability in our country, these requirements are quite varied. Luckily, good quality conventional binders have PG ranges that frequently allow using the same asphalt binder for different climate subzones. Typically, traditional asphalt binders of acceptable quality may have a difference between maximum and minimum temperature of around 86°C. Modified binders may even have greater temperature ranges. For example: typical, non-modified asphalt binders can exhibit the following classifications:

• PG 70-16
• PG 64-22
• PG 58-28
• PG 52-34

In order to determine if a binder is appropriate for a specific climate zone, it must have a high temperature higher than the Txx of the locality, a low temperature lower than the Tyy of the zone and a medium temperature lower than the project site Tint (Kennedy et al., 1994).

 Table 1: PG Grade and Binder Assigned to each Selected Station (I)

The locality of Chile Chico, for example, has an extreme temperature requirement at 50% reliability of PG 52-16 and Tint of 22°C. In this case, it would be appropriate to use the binder PG 58-28 or the binder PG 52-34, which comply with the extreme temperatures and whose medium temperatures are 19°C and 13° respectively. However, it would be appropriate to use the binder PG 64-22, because even if it does fulfill the extreme temperatures, it has a medium temperature of 25°C, which is higher than the Tint of this locality. Consequently, a PG 70-16 should neither be applied.

The “Binder” column of Table 1 shows the binder assigned to each locality in the zoning carried out. The asphalt selection is not necessarily unique for each zone, as we explained earlier. The binders shown in Table 1 were selected so as to keep a relative geographical continuity of the binders recommended, and thus minimize the amount of binders required to cover the entire territory, which favors the implementation and use of the specifications.

Maps for Superpave Binders

In order to make a Superpave zoning based on available temperature data, it is necessary to estimate first the influence area of the weather station. Two criteria were used for this purpose: the climate classification according to the Köppen system for the Chilean territory (Rioseco and Tesser, 2006) as the main criterion and the height above sea level as a complementary criterion. Figure 1 shows the map of Chile according to the Köppen zoning.

Figure 2 shows an example of the Köppen zoning system to delimitate the influence area of the stations available in the area between Coyhaique and Balmaceda on the 50% reliability map. The weather stations closest to Coyhaique required the use of a binder PG 58-28 and they are located within the climate zone classified as Cfc (mild rainy cold without dry season). On the other hand, the weather of Balmaceda indicated the use of a binder PG 52-34 and it is located within the climate zone ET (mild tundra). The limit between both climate zones was used to determine the influence area of each binder, as shown in the Figure.

In the case of stations within a same Köppen climate zone, but with binders of a different PG, their influence zones were determined using the average level curve between the stations. Figure 3 shows the stations Embalse Conchi and Chiu Chiu, which are located within the climate zone BWk’ (cold desert climate), which require a binder PG 58-28 at 98% reliability.

Moreover, the Peine station is located within the same climate zone, with a binder requirement of PG 64-22. Once the heights above sea level of the stations were known (indicated in Table 1), the average level curve was used between adjacent stations in order to delimit the influence area of each PG grade, as shown in the Figure. The same criteria was used for the Caspana station.

Figures 4 and 5 show the results of the Superpave zoning made at 50% and 98% reliabilities respectively. At 50% reliability it was possible to cover almost the entire national territory with three grades of traditional binders: PG 64-22, PG 58-28 and PG 52-34. The sole exception appeared on the foothills of the IX Region, corresponding to the influence area of the stations of Liucura and Lonquimay, which would require a modified binder of PG 64-34.

Figure 1. Köppen Climate Zoning for Chile (Rioseco y Tesser, 2006)


Figure 2. Use of the Köppen Clasiffication in Superpave Zoning (50% Reliability)


Figure 3. Example of the Use of Level Curves for Creating the Chilean Superpave Zoning (98% Reliability)


Figure 4. Superpave Zoning of Chile, 50% Reliability



The same binders, with some changes in the zoning limits, also obtained 98% reliability in most stations. Only the stations of Balmaceda (theoretical PG 58-40 at 98% reliability), Lagunillas (theoretical PG 52-40 at 98% reliability), and Liucura (theoretical PG 64-40 at 98% reliability) did not reach the 98% reliability when the same four binders considered at 50% reliability were used. The reliability achieved in these stations was:

• Balmaceda with PG 52-34: 68% reliability for high and low temperature

• Lagunillas and Liucura with PG 64-34: 98% reliability for high temperature and 68% reliability for low temperature.

Notwithstanding the above, the final selection of the PG grade must consider the characteristics of the project itself, the design effect (loading speed) and/or the heavy traffic volume, which can lead to increase one or two PG grades.

Figure 5. Superpave Zoning of Chile, 98% reliability



5. Conclusions

It was possible to make a Superpave zoning at 50% and 98% reliabilities, based on temperature data from 94 Chilean weather stations belonging to the Chilean Meteorological Service and Chile’s Water Department. The selected stations relied on useful and updated temperature data.

The Köppen climate zoning and the topography of the Chilean territory were useful criteria to define the influence areas of each weather station, which in the end defined the zones recommended for each PG grade.

Despite the variety of PG grades required for different analyzed stations, it was possible to zone most of the Chilean territory, at 50% reliability and three traditional asphalts: PG 64-22, PG 58-28 and PG 52-34. Only a small foothill area in the IX Region required PG 64-34, which will probably be a modified binder.

The same binders, but modifying some of the limits of the influence areas, could be used for a PG zoning at 98% reliability in most stations. However, the stations of Balmaceda, Lagunillas and Liucura achieved a reliability of just 68% when using these same binders. The first two would require PG 52-40 and the third, PG 64-40, in order to reach a 98% reliability.

The resulting Superpave zoning does not constitute a unique solution. In some localities there is more than one traditional binder that can meet the Superpave requirements. The final selection of the binder included practical considerations, such as giving a relative geographical continuity to recommended binders and minimizing the number of binders needed to cover the entire national territory.

The zoning proposed is consistent with the standard condition defined by Superpave, regarding high-speed traffic conditions and moderate traffic volumes. Slow traffic conditions and high traffic volumes should consider increases in high temperature grades.

Original formulas of the Superpave method were used to obtain the temperatures for each locality, thereby yielding more conservative results than those of the LTPP formulas. Future works should consider a comparison using both formulas and recommendations for low-speed traffic and high traffic volumes.


6. Acknowledgements

This study was partially financed by the Research and Postgraduate Department of the Universidad Técnica Federico Santa María through the project CyT-261424. The authors wish to thank the Chilean Meteorological Service for their support during the development of this research and the weather data provided. We are also grateful for the contributions of Mr. Ricardo Alcafuz and Mr. Cristián Díaz in the review of the original work document.

7. References

AASHTO (2015), Standard Practice for Grading or Verifying the Performance Grade (PG) of an Asphalt Binder. R 29-15 444 N Capitol ST. NW-Suite 249 – Washington, DC.

Araya F., A. González, R. Delgadillo, C. Wahr y G. García (2012), Caracterización Reológica Avanzada de Betunes Tradicionales y Modificados Utilizados Actualmente en Chile. Revista Ingeniería de Construcción, Vol. 27, Nº3, pp.198-210.

Asphalt Institute (2003), Performance Graded Asphalt Binder Specification and Testing. Superpave Series N°1 (SP-1), Kentuky.

ASTM (2015), Standard Specification for Performance Graded Asphalt Binder. D6373 West Conshohocken, PA 19428-2959, United States.

Contreras C. (2007), Mapa de Recomendación de Uso de Ligantes Asfalticos Según Clasificación Superpave, Aplicado al Tramo entre Santiago y Los Ángeles. Memoria para optar al Título de Ingeniero Civil, Universidad de Chile.

Delgadillo R., C. Wahr y J.P. Alarcón (2011), Towards the Implementation of the MEPDG in Latin America, Preliminary Work Carried Out in Chile. Transportation Research Record, Vol. 2226, pp. 142-148.

Delgadillo R., C. Wahr, G. García, L. Osorio y O. Salfate (2014), Generating Hourly Climatic Data from Available Weather Information for Pavement Design. Transportation Research Record, Vol. 2433, pp. 48-57.

García G., Marín E., Delgadillo R. y Wahr C. (2014), Caracterización del desempeño a fatiga de mezclas asfálticas mediante los enfoques fenomenológico y de disipación de energía. Carreteras, Vol. 4 (198), pp. 5-14.

Huber G. (1994), Weather Database for the SUPERPAVETM Mix Design System. SHRP-A-648A, National Research Council, Washington.

Kennedy T., G. Huber, E. Harrigan, R. Cominsky, C. Hughes, H. Von Quintus y J. Moulthrop (1994), Superior Performing Asphalt Pavements (Superpave): The Product of the SHRP Asphalt Researh Program. SHRP-A-410, National Research Council, Washington, DC.

Mohseni A. (1998), LTPP Seasonal Asphalt Concrete (AC) Pavement Temperature Models. Federal Highway Administration Report N° FHWA-RD-97-103.

Osorio L., R. Delgadillo y C. Whar (2015), Caracterización y Análisis de la Estadística Chilena para el Diseño de Pavimentos Empírico-Mecanicista. Revista de Ingeniería de Obras Civiles, Vol. 5, pp. 9-17.

Rioseco R. y C. Tesser (2015), Cartografía Interactiva de los climas de Chile. Instituto de Geografía, Pontificia Universidad Católica de Chile,, 2006. Último acceso 15 de marzo de 2015.

Vivanco y Bahia (2005), Transición Hacia un Sistema de Graduación por Desempeño de Betunes Asfalticos para Países en Vías de Desarrollo. XIII Congreso Ibero-Latinoamericano del Asfalto CILA, San José, Costa Rica, Noviembre de 2005.

Universidad Técnica Federico Santa María, Departamento de Obras civiles, Valparaíso, Chile.

Fecha de Recepción: 02/08/2016 Fecha de Aceptación: 12/12/2016



  • There are currently no refbacks.

Copyright (c) 2017 Revista Ingeniería de Construcción

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.