Droughts and erratic rainfall are the primary climate challenges affecting pastoral livelihoods across Kenya's arid and semi-arid lands. This challenge is compounded by a lack of rigorous quantitative evidence on households' long-term adaptive capacity. This study uses Cumulative Adaptation Level (CAL) metric to measure economic resilience by integrating livestock asset retention, adaptation strategy breadth, coping ease, and food security into composite indices. The study draws on a household survey of 490 pastoral households in Kajiado County, Kenya, focusing on the adoption of soil bunding (53.3%) as the focal nature-based intervention. The mean CAL metric (0.605, SD = 0.111) reveals substantial heterogeneity in resilience, which is invisible to single-indicator proxies. Raw comparisons indicate adverse selection, as adopters face higher drought exposure, lower incomes, and greater food insecurity than non-adopters, producing a spurious negative CAL gap of -0.002. Propensity-score matching corrects this bias, yielding an ATT of +0.007, with adopters showing broader adaptation portfolios and fewer months of food insecurity. The economic case for scaling is established through revealed preference. Adopters contribute 92.2 man-days of labour per household, valued at KES[1] 46, 46,094, exceeding estimated construction costs by 2-4 times. Heterogeneity analysis identifies low TLUb and female-headed households as the highest-return adoption targets. The study contributes a replicable CAL framework for measuring dryland resilience and demonstrates that correcting for adverse selection is essential for a credible evaluation of nature-based adaptation investments. [1] KES refers to Kenyan Shilling (Official Kenyan Currency) Exchange rate was 1$US =129.00 at the time of the survey