Pest pressure is a major constraint on crop production in Sub-Saharan Africa, where smallholder farmers rely heavily on synthetic pesticides with potential environmental and health risks. Ecological pest management practices (EPMs) have been proposed as sustainable alternatives, although field-based evidence is limited. This study evaluated pest and natural-enemy responses to EPMs in a cabbage–garden pea rotation system over six years (2020–2025) in Central Kenya. Four treatments were assessed: homemade botanical extracts (Botanicals), coriander intercropping (Intercrop), botanical extracts combined with coriander intercropping (Botanicals + Intercrop), and a control. In cabbage, ecological treatments often recorded lower aphid and diamondback moth abundance and damage than the Control, although these patterns were not consistently significant. Aphid cumulative abundance ranged from an Area Under Disease Progress Curve (AUDPC) of 128.0 under the Control to 47.5 under Botanicals + Intercrop, while DBM abundance ranged from 68.5 to 20.0. In garden pea, aphid damage ranged from an AUDPC of 39.3 under Intercrop to 25.1 under Botanicals + Intercrop, while cutworm damage ranged from 56.3 under Botanicals to 14.8 under Intercrop. Significant treatment effects were detected for selected pest responses. Shannon diversity differed significantly among the treatments in both crop phases. As such, responses were pest- and season-specific rather than uniformly suppressive. These findings identify companion cropping and botanical applications as promising components of EPM within the treatments evaluated. Comparison with conventional pesticide-based or farmer-practice management is however required before recommendations for substitution or wider adoption can be made.

Environmental planning requires a systemic reading of territories where vegetation vigor and exposed soil are treated as complementary indicators of ecological condition and land vulnerability, since approaches based on isolated variables limit both interpretation and action. Guided by this perspective, we evaluated statistical and spatial patterns in vegetation vigor and exposed soil across Kenya’s 47 counties for 2024 and 2025 using Moderate Resolution Imaging Spectroradiometer (MODIS) composites processed in Google Earth Engine. Normalized Difference Vegetation Index (NDVI) from MOD13Q1 and a Bare Soil Index from MCD43A4 were summarized as county means, then examined using Bayesian paired t tests, Bayesian correlations, Bayesian model averaging, random forest clustering, Local Moran’s I, and Besag-York-Mollié 2 model (BYM2) spatial models fitted with INLA. Results show a modest decline in mean NDVI from 0.568 in 2024 to 0.542 in 2025, alongside a shift in mean BSI from -0.023 to -0.007, with Bayes factor evidence for change in both indicators. County rankings were persistent over time for NDVI and BSI, while cross indicator associations were strongly negative, indicating consistent opposition between vegetation cover and soil exposure. Clustering identified three environmental regimes, with a dominant regime characterized by low NDVI and higher exposed soil. Local indicators revealed contiguous clusters of change for both indices, and BYM2 estimates indicated that most latent variance was spatially structured (φ about 0.88 to 0.90). The workflow yields county typologies and maps to guide monitoring and planning, and it establishes a baseline for linking regimes to climate variability and land use trajectories.