Development and Evaluation of a Deep Learning Based System to Predict District-Level Maize Yields in Tanzania

Abstract

Prediction of crop yields is very helpful in ensuring food security, planning harvest management (storage, transport, and labor), and performing market planning. However, in Tanzania, where a majority of the population depends on crop farming as a primary economic activity, the digital tools for predicting crop yields are not yet available, especially at the grass-roots level. In this study, we developed and evaluated Maize Yield Prediction System (MYPS) that uses a short message service (SMS) and the Web to allow rural farmers (via SMS on mobile phones) and government officials (via Web browsers) to predict district-level end-of-season maize yields in Tanzania. The system uses LSTM (Long Short-Term Memory) deep learning models to forecast district-level season-end maize yields from remote sensing data (NDVI on the Terra MODIS satellite) and climate data [maximum temperature, minimum temperature, soil moisture, and precipitation (rainfall)]. The key findings reveal that our unimodal and bimodal deep learning models are very effective in predicting crop yields, achieving mean absolute percentage error (MAPE) scores of 3.656% and 6.648%, respectively, on test (unseen) data. This system will help rural farmers and the government in Tanzania make critical decisions to prevent hunger and plan better harvesting and marketing of crops.

Impact of large kernel size on yield prediction: a case study of corn yield prediction with SEDLA in the U.S. Corn Belt

“…Tende et al used LSTM architecture for predicting corn yield in Tanzania at a district level. They used MODIS NDVI and climate data as inputs to their network and got a MAPE of 3.66 [34]. Although this study has a better performance compared with others it is seen that the monthly average of the lowest and the highest temperature values at the district level do not change much through the evaluation days for the simulation.…”

“…Tende et al used the LSTM model to forecast district-level corn yield in Tanzania. NDVI computed from MODIS, corn crop mask from International Food Policy Research Institute (IFPRI) Spatial Production Allocation Model (SPAM) 2010, and Tmax, Tmean, SM, and PPT from TerraClimate dataset are used as inputs to the proposed LSTM model [34]. Recent research results for yield prediction for a number of crop types through various DL methods are summarized in tables 1 and 2.…”

Rice yield prediction using radar vegetation indices from Sentinel-1 data and multiscale one-dimensional convolutional long- and short-term memory network model

“…The OVIs calculated from MODIS have been more widely used in optical crop yield prediction studies over large regions 9 , 58 – 61 , 68 , 73 …”

Advancements in Utilizing Image-Analysis Technology for Crop-Yield Estimation

“…The spectral information obtained based on remote sensing technology is generally divided into multispectral (MSI) and hyperspectral (HSI) information [29]. The multispectral sensors installed on drones consist of suitable spectral bands in the visible and near-infrared (VNIR) range, which are highly effective in obtaining various vegetation indices (VIs) sensitive to crop health [13], such as the Normalized Difference Vegetation Index (NDVI) [30], Green Normalized Difference Vegetation Index (GNDVI), and Triangle Vegetation Index (TVI), etc. Multispectral data based on drones and combined with machine learning (ML) models [31] has been effectively used to monitor biomass information [32] and yield prediction for various crops, such as corn, wheat, rice, soybeans, cotton, and other varieties.…”