Climate-smart agriculture as disaster risk reduction: Effectiveness varies by hazard type in Pakistan

Climate change increasingly threatens agricultural systems globally, with Pakistan ranking among the most climate-vulnerable countries. This study systematically compared climate-smart agriculture (CSA) effectiveness across different disaster types in Sindh Province, Pakistan: riverine flooding (Dadu), drought (Tharparkar), and coastal flooding (Thatta). A multi-scale analysis combined climate data, area-weighted agricultural vulnerability assessment using 3889 district-crop yield records (2004–2023), and farmer surveys ( n ​= ​88). Climate analysis showed significant Arabian Sea warming (0.41 ​°C ​decade −1 , p ​< ​0.01) with high precipitation variability (coefficient of variation ​≈ ​72 ​%). The corrected area-weighted vulnerability assessment ranked Tharparkar most vulnerable (#1, index: 66.23), followed by Thatta (#8, 31.69) and Dadu (#19, 24.77) among 23 districts. Drought analysis identified 2 of 20 years meeting drought criteria (Standardized Precipitation Index ≤ −0.5) in Tharparkar. CSA adoption varied significantly by disaster type (Kruskal-Wallis H ​= ​74.06, p ​< ​0.001). Strong positive correlation between CSA practices and disaster resilience emerged in drought-vulnerable Tharparkar ( r ​= ​0.756, p ​< ​0.001, explaining 57 ​% of variance), while flood-prone districts showed negligible relationships (Thatta: r ​= ​−0.089, p ​> ​0.05; Dadu: zero variance). Future projections indicate substantial sea-level rise of ∼331–751 ​mm by 2100 across Shared Socioeconomic Pathway scenarios. Results demonstrate CSA practices effectively reduce climate disaster risks in drought-prone systems, but effectiveness varies by disaster type, requiring tailored implementation approaches. This research provides the first systematic evidence comparing CSA effectiveness across different disaster contexts, supporting targeted rather than uniform adaptation policies. • CSA effectiveness varies by disaster type: strong in drought, negligible in flooding. • Arabian Sea warming accelerated fourfold to 0.41 ​°C ​decade −1 since 2004. • Drought-prone Tharparkar ranked most vulnerable among 23 districts analyzed. • CSA practices explained 57 ​% resilience variance in drought-vulnerable areas. • Targeted adaptation policies needed rather than uniform CSA approaches.