Small-scale farming optimization using frugal plant-based irrigation scheduling in Kenya

Abstract:

Climate change is altering environmental conditions and impacting agriculture globally, especially in sub-Saharan Africa. Increased severity and duration of droughts coupled with decreased rainfall, mean that farmers have smaller inconsistent water supplies for crop production. Yields are negatively impacted by both deficiencies and excesses in key nutrients, as well as water in the surrounding environment. Small-scale farmers either overlook these nuances, make adjustments based on guesses as field conditions evolve or rely on confusing advice from agronomists and technical experts, who also lack insights in the farmer’s unique operating situation. Thus, their crop yields are often below optimal quantities as farmers are unable to ensure that their crops experience limited stress during development. Precision irrigation scheduling was designed to address this need but is expensive and does not match the mode of operation of the average small-scale farmer. To this end, we have developed a frugal, cloud-integrated, cyber-physical sensing framework which relies on small networks of strategically placed soil moisture sensors together with tank-level sensors and utilizes plant growth models to determine if and how much irrigation is needed. By utilizing weather data to calculate both plant growth and evapotranspiration, we are able to monitor plant health and track development. Combining these calculations with cloud-enabled data management and analytics, we seek to provide small-scale farmers with a low-cost method for precision irrigation management. From both laboratory trials and a case study (a mixed-crop farm in Machakos), we illustrate the effectiveness of our approach for validating the accuracy of sensors and developing models and algorithms for tailored irrigation scheduling.