Conventional approach of establishing soil conservation strategies in degraded drylands has had negligible success. This has been contributed by many constraints, including; lagging of farmers in technology adoption, inadequate resources, and lack of motivation. Thus, a study was conducted among three agro-pastoral community farmer groups in Korellach Parak, Kapkitony, and Kaporowo villages domiciled in Chepareria ward, West Pokot, Kenya, to assess contributory factors and consequences of adopting terracing as a soil conservation measure. Mixed methods comprising; one-on-one interviews, cross-sectional field measurements, and focus group discussions (FGDs), were used for data collection. Results indicate that the agro-pastoral communities are fully aware of soil degradation and its impacts. Besides terracing, farmers practice stone bands, enclosures, agroforestry, and ridges. Terracing is a recently adopted farm-level soil conservation practice achieved through organized farmer groups dubbed “Kemorokorenyo” (meaning let us reclaim our land) merry-go-round. Within the three villages, 60% of the households have their farms terraced with an average terrace volume of 103.8±21.45m3, 105.89±33.126m3, and 129.6±15.966m3 in Parak Kapkitony and Kaporowo, respectively. Rapid sedimentation of terraces dykes, which contributes to the reduced effectiveness of the terrace system was identified as the major challenge. The sediment volume significantly differs along the slope, with the highest sediment build-up experienced on high slopes as shown by the Kruskal Wallis test; H (2) =6.699, p=0.035. Terrace embankments reinforcement practice to counter sedimentation challenge has faced slow adoption. The poor reinforcement is attributable to the lack of knowledge on suitable local context multipurpose materials to meet the community’s needs.

Soil erosion threatens the sustainable intensification of food systems among smallholder farmers in arid and semi-arid lands (ASALs). Intensifying adoption of soil mitigation and rehabilitation measures is thus needed urgently in these ASALs, but scaling up these measures depends on scientific evidence of their contributions to key components of sustainable intensification such as soil organic carbon. However, there is no information on how existing mitigation and rehabilitation measures influence soil carbon fractions and carbon management indices in ASALs. This study evaluated the influence of soil erosion mitigation and rehabilitation measures on soil carbon fractions and management indices in Arenic Lixisols of semi-arid environments in West Pokot County, Kenya. We evaluated different vegetation types (maize-beans intercrop and pastures) with and without two locally developed terrace designs for soil conservation (Fanya Juu and Fanya Chini). Combining terracing with annual cropping significantly increased total organic carbon (TOC). The highest TOC (13 g C kg−1) was recorded in pasturelands with terraces while degraded land with no intervention was found to have the lowest TOC (6.0 g C kg−1). Terraced farms with longer residence time (>4 years old) had significantly higher organic carbon than (<4 years old). Other soil properties remained stable with terrace age (1–5 years). Labile SOC and non-labile SOC differed significantly within and across vegetation types with or without terraces (p < 0.05). Pasture and crop systems with terraces had high labile SOC content of 5.9 g C kg−1 and 7.2 g C kg−1, respectively. Labile SOC followed the TOC trend with terrace age, i.e., increasing from 1 year to 5 years old. Combined pasture and terraces had a significantly higher carbon management index (CMI) of 161.7, or 14 times the CMI found in degraded systems with no interventions and 1.5 times the combined crop system with terraces. CMI was also directly correlated with residence time terraces had stayed in the crop system, increasing from 1 year to 5 years old. Contrary to CMI and other indices, the weighted enrichment ratio was found to inversely correlate with age of terrace. Improvement of carbon content and CMI resulted from restorative measures and likely improved soil quality and ecosystem functions. Although terraces play a significant role in the restoration of degraded soils as indicated by the above-mentioned changes, they are most beneficial when used in combination with croplands because of the high level of disturbance and flows of both inputs and outputs of carbon for these croplands.

The inadequacy of information on rice farmers' situations, particularly their access to propagation materials and grain storage and the impact thereof on Liberia's food security scenario, curtails the development of strategies and interventions meant to optimize rice production. This study investigated the rice seed systems, storage methods of the harvested rice grains, and degree of fungal contamination of stored rice seeds in the major rice-producing counties of Liberia: Lofa, Bong, Montserrado, and Nimba. A mixed data collection method, comprising interviews and focus group discussions (FGDs) was adopted. Five hundred (500) farmers were purposefully selected for one-on-one interviews, and 12 FGDs were held (three in each county). The results indicated that 94.7% of farmers source seeds through informal channels. Grain for use as seed in the subsequent farming season is mainly stored in kitchen attics, a practice reported by 83.8% of the farmers, while 7.8%, 3.8%, and 4.6% of farmers stored seeds in plastic containers, nylon sacks, and jute bags, respectively. Land size was identified as the primary factor determining rice yield across the studied counties, R2 = 0.944, p = 0.001. Farmers in high-rainfall regions had a high likelihood of experiencing fungal infections on their stored grains; however, only 19.6% of farmers were aware of the health implications of consuming affected grains. Therefore, policies and support frameworks should be directed towards actualizing modern seed channels and extension services and creating awareness of the different nodes of the rice value chain.

This study aimed at (i) developing, evaluating and comparing the performance of support vector machines (SVM), boosted regression trees (BRT), random forest (RF) and logistic regression (LR) models in mapping gully erosion susceptibility, and (ii) determining the important gully erosion conditioning factors (GECFs) in a Kenyan semi-arid landscape. A total of 431 geo-referenced gully erosion points were gathered through a field survey and visual interpretation of high-resolution satellite imagery on Google Earth, while 24 raster-based GECFs were retrieved from the existing geodatabases for spatial modeling and prediction. The resultant models exhibited excellent performance, although the machine learners outperformed the benchmark LR technique. Specifically, the RF and BRT models returned the highest area under the receiver operating characteristic curve (AUC = 0.89 each) and overall accuracy (OA = 80.2%; 79.7%, respectively), followed by the SVM and LR models (AUC = 0.86; 0.85 & OA = 79.1%; 79.6%, respectively). In addition, the importance of the GECFs varied among the models. The best-performing RF model ranked the distance to a stream, drainage density and valley depth as the three most important GECFs in the region. The output gully erosion susceptibility maps can support the efficient allocation of resources for sustainable land management in the area.