Selecting farms and mapping climate regions

First, we developed various datasets of cacao farms occurrences. In the sampling description methods, more details are provided about the farms' selection process. Then, the following step involved extracting the values of environmental variables that corresponded to the geographical locations of cacao farms. Next, an online statistics calculator (https://www.statskingdom.com/pca-calculator.html) was used to generate a standard Principal Component Analysis (PCA) separately for the macro and microclimate variables. The loading values of PCA Axis 1 were utilised to create a gridded map of macro- and microclimate regions using the QGIS software (QGIS 2019). These geospatial methods enabled us to identify the Arauca cacao regions through a national scale analysis.

Results

National and regional scale

We identified two national and six regional climate regions for cacao in Arauca (Fig. 1). High temperatures and low rainfall are reported on the east of the region creating a clear east to west climate pattern, as shown in Fig. 1C-D. The Arauquita Region in the northeast is hotter than the western extremes of Saravena. The foothills of Saravena, Fortul and Tame have cooler and wetter conditions. A transition zone is observed in the central region of Arauca, with mixed conditions shared by the boundaries of Saravena, Fortul and Tame. The southeast of the region, mainly covering areas of Tame, showed different conditions from the rest of the region, with a predominance of maximum temperature during the dry season and low precipitation during the wet season as key variables (Fig. 1B). A lower degree of spatial grouping was recovered, based on soil temperature regions (Fig. 1C). However, a northeast and west group of similar soil temperature conditions in Saravena and Arauquita was identified, likely influenced by the specific conditions of the Arauca riverbanks. The foothills of Fortul and Tame show similar conditions, whereas Tame seems to follow the river lines with a clear east to west gradient. Temperature seasonality shows a higher percentage of influence on the northern areas of Arauquita and Saravena, suggesting that soil temperature conditions are more variable in these areas than in the southwest foothills of Tame and Fortúl. The eastern part of Tame also seems to have a high degree of variance explained by temperature seasonality (Fig. 1D).

Figure 1.  

Maps of cacao climate regions in Arauca, Colombia. National (a-b) and regional (c-d) scale maps of cacao farms according to climate and soil temperature in Arauca Department, Colombia. Blue lines are rivers and red lines roads. Blue dots are indicative of negative values of the PCA, cream and red dots are indicative of positive PCA values. Cream to orange dots are indicative of mixed conditions.

Local scale

We identified nine clusters comprising 180 farms (Fig. 2). Based on the specific climate clusters (Fig. 2A), it was possible to identify the highest minimum and average temperature, wind velocity and solar radiation in clusters 2 and 3, located in the northeast of the study area. In contrast, the lowest values of these variables were identified in clusters 4, 5 and 7, which are in the northwest and southwest of the study area. Regarding rainfall, during the dry season, the lowest volumes were identified in clusters 1, 2, 3 and 9 (mainly in the west), while during the rainy season, the highest values were found in clusters 1, 6, 7, 8 and 9, in the central study area. The results of the nine soil clusters (Fig. 2B) revealed that, in the topsoil, there are clusters associated with riverbanks following floodplains (Clusters 1 to 5) and non-flooding foothills or rough plains (Clusters 6 to 9). We also considered other critical factors or cacao crop, such as the presence of cadmium in both soils and cacao beans (Bravo 2022).

Figure 2.  

Maps of climate clusters in Arauca, Colombia. Climate (A) and soils temperature (B) clusters at local scale across the sampling area of 180 cacao farms in Arauca.

A national scale mapping was conducted using a previously-published occurrences dataset comprising 3,141 cacao farms in Colombia (González-Orozco and Pesca 2022). To establish the climate regions, a spatial analysis method derived from González-Orozco et al. (2014) was applied to the selected national cacao farms dataset. Seven macroclimate (Fick and Hijmans 2017) and eleven microclimate soil temperature variables at a depth of 0-5 cm were used to characterise the agrobiodiversity and the climate zones (Lembrechts 2022) of the cacao regions in Arauca. The spatial resolution of the environmental variables was 1 km².

This method involved extracting the values of environmental variables that corresponded to the geographical locations of cacao farms. Next, a Principal Component Analysis (PCA) was separately conducted for the macro- and microclimate variables. The loading values of PCA Axis 1 were utilised to create a gridded map of macro- and microclimate regions using the QGIS software (QGIS 2019). These geospatial methods enabled us to identify the Arauca cacao regions through a national scale analysis.

At the regional scale, a dataset of 1,538 cacao farms in Arauca was developed (regional scale dataset). Eleven local agronomy professionals visited cacao farms in the Municipalities of Saravena, Arauquita, Tame and Fortul in Arauca. Each farm was georeferenced and uploaded to a GIS desktop. Then, the González-Orozco et al. (2014) climate mapping method was applied to the cacao farms dataset. The analysed climate variables were subdivided into temporal groups representing the dry (Nov-Feb) and wet (Mar-Oct) seasons that Arauca experiences throughout the year. The grouped macroclimate and soil temperature databases and all farms were used to map the climate and soil temperature regions using a PCA spatial mapping method (González-Orozco et al. 2014).

At the local scale, a combined effort amongst cacao farmers, agronomy field workers and a multidisciplinary team of researchers pre-selected a subset of 180 cacao farms as the best cases for future sampling efforts of cacao and other environmental variables from the 1,538 farms. The selected farms were chosen to represent good variability in climate and soil temperature conditions (Fick and Hijmans 2017, Lembrechts 2022), cadmium soil hotspots (Bravo et al. 2021), landscape types, cacao varieties and healthy cacao plantations. Other parameters, such as proximity to the road network, were also considered. Additionally, the criteria for chosing the local-scale farms included agronomic, socioeconomic, territorial and environmental traits. A second-step process was applied to classify the 180 local-scale farms selected into spatial clusters, resulting in the identification of nine clusters that were used as the sampling regions (Local-scale dataset).

A presence-only matrix of tree species (Farm-scale dataset) associated with cacao was presented as part of the agrobiodiversity assessment done across the Arauca climate zones.

Department of Arauca, Colombia. Municipalities: Saravena, Fortul, Tame and Arauquita.

6.97 and 7.00 Latitude; -71.76 and -71.61 Longitude.

Creative Commons Public Domain Waiver (CC-Zero)