Tropentag, September 16 - 18, 2026, Göttingen "Towards multi-functional agro-ecosystems promoting climate-resilient futures"

Spatial gaps and maladaptation risk: performance frontiers for climate resilience in Kenya

¹Leibniz-Centre for Agricultural Landscape Research (ZALF), Development-oriented International Agricultural Research (DIA), Germany
²Leibniz-Centre for Agricultural Landscape Research (ZALF), Development-oriented International Agricultural Research, Germany
³Humboldt-Universität zu Berlin, Thaer-Institute of Agricultural and Horticultural Sci., Germany

Abstract

Climate-smart agriculture (CSA) is widely promoted to build resilience, yet adaptation benefits remain unevenly distributed and maladaptation risk – where households become trapped in low-yield, high-gap states – is poorly understood. Using survey data from 569 smallholder households in Western Kenya, stratified by gender and agro‑ecological zone (Lower Midland Zone LMZ, Upper Midland Zone UMZ), we diagnose structural performance gaps and spatial patterns of maladaptation risk. We apply non‑parametric conditional quantile frontiers (τ=0.90) to benchmark attainable maize yield given observed adaptation inputs (crop diversification, soil and water conservation). The distance‑to‑frontier measures how far households are from feasible performance – a diagnostic of structural inefficiency. Maladaptation risk is defined as a household simultaneously in the lowest yield quartile (≤25^th percentile) and the highest distance‑to‑frontier quartile (≥75^th percentile). Spatial clustering is assessed using global Moran’s I and local indicators of spatial association (LISA).

Results show stark agro‑ecological inequality. Mean distance‑to‑frontier is 0.172 t/ha in the LMZ versus 0.061 t/ha in the UMZ – a 2.8‑fold larger gap in the more marginal zone. Maladaptation risk is 0% in the UMZ but 15% in the LMZ (16% for male‑headed, 12% for female‑headed households). Global Moran’s I for frontier distance is 0.445 (p=0.002), confirming significant positive spatial clustering: disadvantage is not random but geographically embedded. A policy simulation of gap‑based targeting reduces mean inefficiency by 7.4% but, critically, increases within‑group inequality (ΔGini +0.010 to +0.013). This reveals an equity‑efficiency trade‑off: focusing only on the largest performance gaps can inadvertently widen disparities.

We conclude that performance gaps are place‑based and that maladaptation risk concentrates in geographic hotspots. Equity‑calibrated, spatially targeted interventions – such as hotspot‑guided extension, infrastructure investment, and tailored insurance – are essential for building truly multifunctional and climate‑resilient agro‑ecosystems. The framework is portable and can be adapted to other regions facing similar structural constraints.

Keywords: Agro-ecological zones, climate-smart agriculture, conditional quantile frontiers, gender, Kenya, lisa, maladaptation risk, moran's i, performance frontiers, smallholder agriculture, spatial clustering

Contact Address: Denis Momanyi, Leibniz-Centre for Agricultural Landscape Research (ZALF), Development-oriented International Agricultural Research (DIA), Eberswalder str. 84, 15374 Müncheberg, Germany, e-mail: denis.momanyizalf.de