Soil stewardship to climate proof Ghana’s cocoa sector
Mainstreaming climate smart practices in cocoa in Ghana is the declared objective of our work. The differentiated nature of the risks from climate change requires differentiated strategies for managing adaptation. The challenge is to support the large-scale adoption of practices that are climate smart. This project seeks to develop incentives and support mechanisms that will drive farmer uptake of CSA at scale.
The first step is to put the project onto solid science based feet to identify different climate change impact zones in Ghana. From this differentiated zoning climate hazards will be described to tailor CSA practices to address the local challenges. Preliminary results were presented at a kickoff workshop in Accra in May to a wide multi-stakeholder panel. These first results were received with overall agreement by the local experts but we were sent back to our offices with some homework to do: think about the role of soils, clearer climate indicator variables and inclusion of protected areas.
What did we learn in High School? When you have tough questions to answer – ask the guy who already read the book. Facilitated by our partners in Ghana (Sander Muilerman (IITA) and Christan Mensah (Rainforest Alliance) we teamed up with our colleagues at the Cocoa Research Institute of Ghana (CRIG) for a one week workshop to improve our climate change impact gradient mapping and to triangulate CSA practices identified during field surveys. Under the leadership of Dr. Amos Quaye a group of ten of their cocoa researchers had CIAT’s Christian Bunn run und re-run his models over and over again until we were finally able to meet their ground-truthed expectations. In between model validation session a committed debate evolved over the viability of the suggested list of CSA practices presented by Patrick Adjewodah (IITA/RA) (Watch this space for more on this to come!).
The final results of this process are exciting to say the least. Our usual methodology uses machine-learning algorithms to estimate the climatic suitability of a crop. A classifier such as RandomForests (RF) is taught to separate occurrence locations from a random background sample using climate data and then used to make predictions on climate maps. What you want to get right is to have all potentially important variables included and that the occurrence data is representative of the complete distributional range. This is where CRIG’s support really showed to be very valuable.
Our first results had underestimated some cocoa area at the margins for historic climate conditions. Under future conditions the model projected increases in suitable area where cocoa will never be grown because of the unfit soils. CRIG’s input helped to identify climate variables that are well suited to define climatic zones of cocoa production, to stratify the occurrence dataset for cocoa that we used and to work out a way to include soil data for our suitability mapping. The soil scientists chose a set of 40 soil variables that describe soil characteristics which may provide resilience to climate hazards – such as drought – to the cocoa tree. Soil organic matter, rootability, silt, sand, and clay content, and exchangeable bases, acidity, cations went as variables into a RF suitability classification model.
The result soil suitability map surprised us. It is a very accurate descriptor of the spatial distribution of cocoa production locations. Of course, the soils in Ghana were shaped by the region’s vegetation that evolved in a humid climate. Nevertheless, it was interesting to see that using only climate data our suitability model would underestimate some isolated cocoa producers that are spread over the savanna zone. The soil suitability map correctly predicted these areas as potentially useful. Albeit, as CRIG’s experts stated, these are exceptional farmers that had to create innovative approaches to cultivate cocoa in their sub-optimal climate.
But it is exactly this kind of lesson that is needed to understand viable farmer driven approaches to climate change adaptation. What we see in Ghana is that the area that in the future may become climatically unsuitable – if nothing is done – will still be useful if we are able to learn the right lessons. We now know where to look, who to talk to, and which threats have to be addressed to make Ghanaian cocoa climate resilient. Keeping the soils fertile, to protect them from erosion or deterioration of organic matter will be one of the most important strategies to make cocoa production in Ghana climate smart.