Integration of Indigenous Knowledge, Climate Data, Satellite Imagery and Machine Learning to Optimize Cropping Decisions by Small-Scale Farmers. a Case Study of uMgungundlovu District Municipality, South Africa

Abstract

Despite being the backbone of food security in most African countries, small-scale farmers are overlooked during the implementation of many of developmental projects. These farmers are financially incapable of equipping themselves with irrigation systems and other agricultural technologies that can assist in improving their farm yields. One of the challenges facing them is that they rely heavily on rain-fed agriculture which makes them extremely vulnerable in the face of climate change. They have continued to consult their indigenous knowledge systems to predict the onset of rains and in making critical decisions such as when to prepare land for crop cultivation. Evidence shows that this knowledge is no longer as precise as it used to be – among other reasons, this is due to the effects of climate change and deforestation. The second problem is that the only sources of weather information these farmers have (e.g. the media) are general and not scaled down to the specific locations where they reside. On the other hand, most of the small-scale farmers are educationally and technologically semi-literate and are financially crawling when it comes to adoption of the likes of sensors and other technologies that could help in predicting and monitoring crop health. There is however opportunity in that most of these farmers are now using android phones. In this research, a model that utilizes the indigenous knowledge, climate data and vegetation index to foresee the onset of the favourable weather season for crop cultivation, crop monitoring and crop health prediction is proposed.