Accurate livestock counts are essential for effective pastureland management. High spatial resolution remote sensing, coupled with deep learning, has shown promising results in livestock detection. However, challenges persist, particularly when the targets are small and in a heterogeneous environment, such as those in African rangelands. This study evaluated nine state-of-the-art object detection models, four variants each from YOLOv5 and YOLOv8, and Faster R-CNN, for detecting cattle in 10 cm resolution aerial RGB imagery in Kenya. The experiment involved 1039 images with 9641 labels for training from sites with varying land cover characteristics. The trained models were evaluated on 277 images and 2642 labels in the test dataset, and their performance was compared using Precision, Recall, and Average Precision (AP_0.5–0.95). The results indicated that reduced spatial resolution, dense shrub cover, and shadows diminish the model’s ability to distinguish cattle from the background. The YOLOv8m architecture achieved the best AP_0.5–0.95 accuracy of 39.6% with Precision and Recall of 91.0% and 83.4%, respectively. Despite its superior performance, YOLOv8m had the highest counting error of −8%. By contrast, YOLOv5m with AP_0.5–0.95 of 39.3% attained the most accurate cattle count with RMSE of 1.3 and R² of 0.98 for variable cattle herd densities. These results highlight that a model with high AP_0.5–0.95 detection accuracy may struggle with counting cattle accurately. Nevertheless, these findings suggest the potential to upscale aerial-imagery-trained object detection models to satellite imagery for conducting cattle censuses over large areas. In addition, accurate cattle counts will support sustainable pastureland management by ensuring stock numbers do not exceed the forage available for grazing, thereby mitigating overgrazing.

Global climate change is projected to disproportionately impact cereal crop yields in developing regions, such as Kenya, due to increased vulnerability and limited adaptation capacity of the population. This study examines the current and projected influence of climate change on maize yields in two major maize-producing counties of Kenya. Utilizing the calibrated and evaluated DSSAT-CERES-Maize model (where DSSAT is Decision Support Systems for Agrotechnology Transfer and CERES stands for Crop and Environment REsource Synthesis) for the H614 maize cultivar, we investigated the projected impact of climate change on maize production with reference to a baseline period (1984–2013). Simulations were conducted for the mid-century period (2041–2070) and end-of-century period (2071–2100) using projected climate data from regional climate models (RCMs) under two Representative Concentration Pathways (RCPs; 4.5 and 8.5) scenarios. Our findings indicate a substantial decline in maize yields, ranging from 7 to 20% for the mid-century period and between 22 and 41% for the end-of-century period, with increased temperature during critical growth phases identified as the primary driver. Spatial clustering and hotspot analysis reveal differential climate impacts across the region. In the end-of-century period, both scenarios revealed that the counties will be marked by hotspots and adaptation spots, areas where climate change adaptation should be intensified. The study underscores the urgency for tailored, location-specific adaptation measures such as maize-legume intercropping, drought-resistant crops, soil water conservation and optimum sowing to mitigate future yield losses and adapt maize production to climate change.

Climate change is expected to significantly affect agricultural production in East Africa (EA). In this study, we synthesized the DSSAT-CERES-Maize model calibrated and evaluated experiments to analyze the sensitivity of climatic variables on maize yield in the region. We used calibrated cultivar coefficients of locally adopted varieties in twelve sites across the region. Consequently, we generated synthetic scenarios of precipitation and temperature changes in line with the plausible projections of the Intergovernmental Panel on Climate Change (IPCC) to characterize the impact of climate change on maize production across the region. Our findings reveal that the impacts of climate change are heterogeneous and vary from location to location. The analysis points to adverse effects in the semi-arid zones, with maize production in Katumani (Kenya), Dodoma (Tanzania), and Ruzizi (Rwanda) expected to decline by -25% to -30% under an extreme temperature rise of + 3 °C and a 30% decline in precipitation. The results also reveal that increased precipitation will compensate for yield losses resulting from elevated temperatures in both arid and humid zones. The potential yield gain under increased precipitation and warming is 16%, 18%, and 5% in Katumani, Dodoma, and Morogoro (Tanzania), respectively. The study recommends site-specific soil, water, and land management adaptation strategies. Strategies for soil and water conservation are recommended for dry regions, whereas approaches such as varying sowing dates are recommended for semi-humid to humid zones. Nutrient enhancement and cultivar variation might be feasible in both contexts.

Maize (Zea mays L.) productivity in Kenya has witnessed a decline attributed to the effects of climate change and biophysical constraints. The assessment of agronomic practices across agroecological zones (AEZs) is limited by inadequate data quality, hindering a precise evaluation of maize yield on a large scale. In this study, we employed the DSSAT‐CERES‐Maize crop model (where CERES is Crop Environment Resource Synthesis and DSSAT is Decision Support System for Agrotechnology Transfer) to investigate the impacts of different agronomic practices on maize yield across different AEZs in two counties of Kenya. The model was calibrated and evaluated with observed grain yield, biomass, leaf area index, phenology, and soil water content from 2‐year experiments. Remote sensing (RS) images derived from the Sentinel‐2 satellite were integrated to delineate maize areas, and the resulting information was merged with DSSAT‐CERES‐Maize yield simulations. This facilitated a comprehensive quantification of various agronomic measures at pixel scales. Evaluation of agronomic measures revealed that sowing dates and cultivar types significantly influenced maize yield across the AEZs. Notably, AEZ II and AEZ III exhibited elevated yields when implementing combined practices of early sowing and cultivar H614. The impacts of optimal management practices varied across the AEZs, resulting in yield increases of 81, 115, and 202 kg ha⁻¹ in AEZ I, AEZ II, and AEZ III, respectively. This study underscores the potential of the CERES‐Maize model and high‐resolution RS data in estimating production at larger scales. Furthermore, this integrated approach holds promise for supporting agricultural decision‐making and designing optimal strategies to enhance productivity while accounting for site‐specific conditions.

Land Use and Land Cover (LULC) changes due to human activities and natural factors significantly influence the variation in Ecosystem Service Values (ESVs). This study assessed the spatio-temporal dynamics of ESVs in Baringo, an Arid and Semi-Arid Land (ASAL) county located in Kenya’s Great Rift Valley, over 24 years (2000–2024). The Benefit Transfer Method (BTM) was employed to estimate ESV changes using LULC data for the years 2000, 2014, and 2024, alongside updated valuation coefficients from the Ecosystem Services Valuation Database (ESVD) published by de Groot and others in 2020. The analysis revealed a substantial decline in the total ESVs in Baringo County, from United States Dollars (US$) 25.70 billion in 2000 to US$14.92 billion in 2024, representing a loss of US$ -10.78 billion (-41.9%). This translates to an average annual ESV loss of approximately US$ -448.96 million. Forestland reduction emerged as the primary driver of this decline, whereas expansion of water bodies contributed positively to the total ESV in the study region. These findings underscore the urgent need to combat deforestation and forest degradation while promoting the sustainable management of water resources. Such measures are essential for reversing ESV losses and achieving key Sustainable Development Goals (SDGs), particularly SDG 6 (Clean Water and Sanitation), SDG 13 (Climate Action), and SDG 15 (Life on Land).