Validating Terra-i in the Peruvian Amazon
During the 1st and 12th of June 2015, the Terra-i team, together with the Instituto de Investigaciones de la Amazonía Peruana (IIAP) and the Universidad Nacional Agraria La Molina (VLIR-UNALM), conducted the second field validation of the data produced by the Terra-I system. This time, the study area was the Yurimaguas district, Alto Amazonas province, Loreto region (Peru).
We used information on populated places, main roads, rivers and information on land cover changes detected for 2013, 2014 and 2015 to define the 65 sampling points (or Terra-I pixels) for the validation process (Figure 1).
We visited these selected points by road, river and even air (yes, we flew over a part of the study area). We started by road dividing our group in two: the first group composed by Reynaldo Solis from IIAP, Paula Paz from the Terra-I team, and Pedro Ramirez, and the second group composed by Ruben Coppus from CIAT, Jhon Tello from the Terra-I team, and Eduardo Fasabi. On the third day of our visit, the group composed by Ruben and colleagues started a trip of four days along the Huallaga River. The last three days of our field work and under the framework of the technological and knowledge transfer agreement between the DIGOT-MINAM and Terra-I CIAT, we had Witman Garcia and Eloy Valencia accompanying our activities. Additionally, in order to reach various sampling points with difficult access by road or river, we took a small plane and flew over the area for one hour approximately. During our trip we could enjoy the beauty of the landscapes, the local biodiversity, good food, and friendly and generous people always keen to help (Figure 2).
The selected study area is in the Yurimaguas district (Alto Amazonas province, Loreto region, Peru). The land cover loss of the district for the period between 2004 and April 2015 is 25869 Ha, with an annual rate of 2156 Ha/year according to the measurements obtained via the Terra-I system. We used the data for years 2013, 2014, and 2015 for the field validation, finding 5081 ha of land cover loss detected through Terra-I (Figure 3).
The economy in Yurimaguas is based in agriculture, livestock raising and forestry. The major products from the region are: palm oil, rice, cassava, yellow maize and starchy bananas (or plantains). The most extended activity is subsistence agriculture, however palm oil cultivation has been dramatically growing during the last couple of years, reaching an extension of 11662 ha and a production of 107230 tonnes for 2014. Likewise, 870 Has have been established with cacao during the last four years. Livestock raising is focused in cattle ranching, porcine husbandry and poultry, mainly.
Three buffer areas of 1, 3, and 5 km were estimated surrounding roads and main rivers in the region (Figure 1). We defined three paths to visit: 1) using the road that connects Yurimaguas with the town of Munichis, 2) following the road to the town of Grau, and 3) throughout the Huallaga River. We measured the density of Terra-I detection pixels in the region for each year, and for each buffered distance, and afterwards randomly selected the pixels for the field validation, to guarantee the criteria of giving all pixels the same probability of being selected (Olofsson et al., 2014).
Based on the methodology on best practices for land use change studies from IPCC (2013), we estimated a subsample of nine points within each pixel, assuming that a detected change might only occur in a partial section of the pixel (a Terra-I pixel measures 250 x 250 meters, equivalent to 6.25 Ha) (Coca et al., 2014) (Figure 5).
We took notes from each of the visited sites, describing whether there are or not changes in the land use, the current type of land cover, species found, and current land use. Moreover, we took pictures on each sub-sampling site on the four cardinal directions (N, E, S, W). We also surveyed local participants at each sampling area when possible.
According to our observations, we could detect a pattern of accelerated land use change in the areas close to roads due to the growing number of paved roads and secondary pathways. Most of the changes near roads are associated to palm oil cultivation, even reaching an extension above 2000 Has in the case of plantations owned by private companies. Other crops and livestock for milk and meat were identified in smaller extensions owned by local farmers. Milk and meat are locally sold in Yurimaguas and Munichis.
We also identified patches of 40 to 50 Has approx. in pastures and multiple rice plantations of 30 to 50 has, located in the floodplains of the rivers, and close to the commercial routes to Yurimaguas, Iquitos and San Martin. Small cultivations of palm heart, maize, plantain, cassava and peanut were also found in smaller areas.
Cultivations of papaya of about 20 to 40 Has were recorded over the Huallaga River, close to the populated places of Mondongo, Cachihuañusca, Rosa de América, San José de Zapote, Santa Isabel, Jeberillos and 4 de octubre. The papayas produced at these places are sent and commercialized in Lima, Chiclayo and Chimbote. Pastures of 20 – 40 Has with livestock for milk and meat were also seen throughout the edges of the river, and the commercialization of the products derived from this activity takes place locally in Yurimaguas and Iquitos. Plantains and cassava grown in this area are for local consumption.
Finally, during our flight over the study area, we reached sampling points that were not accessible by road neither rivers, most of these sampling points are found within palm oil cultivations (Figure 6).
The population growth in the area of study has created an increasing demand of land for subsistence agriculture, and more recently for commercial agriculture. Areas highly accessible by roads and rivers are being sold by smallholders to large private companies, or to switch their current crops into high value crops, as it is the case of palm oil. Relicts of primary forest are still found within these areas, and under great threat of disappearing.
The field validation of Terra-i data helps us to calibrate our methodologies and therefore improve the precision of our results. This is translated into better decision making for protecting our forests.
We acknowledge all the people that made possible this field work.
Olofsson, P., Foody, G. M., Herold, M., Stehman, S. V., Woodcock, C. E. & M. A. Wulder. 2014. Good practices for estimating area and assessing accuracy of land change. Remote Sensing of Environment 148, 42–57.
Coca-Castro, A. 2014. Reporte de la validación en campo del Sistema Terra-i (CIAT) en las comunidades del proyecto ASSETS en la región de Ucayali, Perú. 53 P. En: http://goo.gl/noQsYd
IPCC – Intergovernmental Panel on Climate Change. 2003. Good Practice Guidance for Land Use, Land-Use Change and Forestry. 590 p.
Check more pictures of our field work here: http://www.flickr.com/photos/131075783@N02/albums
Written by: Jhon Jairo Tello, Paula Andrea Paz y Reynaldo Solis. This work was undertaken as part of the project “Sustainable Amazonian Landscapes” which is supported by the International Climate Initiative (IKI). The project is led by CIAT and implemented together with the Potsdam Institute for Climate Impact Research (PIK), the Center for Research on Sustainable Systems of Agriculture Production (CIPAV), Instituto Amazónico de Investigaciones Científicas SINCHI, Universidad de la Amazonía, Instituto de Investigaciones de la Amazonía Peruana (IIAP), and Universidad Nacional Agraria La Molina (VLIR-UNALM).
English version prepared with help of Nora Castañeda (CIAT).