Evaluating the role of trees in croplands in mitigating the health effects of extreme heat exposure among agricultural workers in Tanzania: a protocol for a quasi-experimental study.

<ns5:p>Background Climate change is intensifying occupational heat exposure globally. Among the most vulnerable populations are agricultural workers in semi-arid sub-Saharan Africa. Recurrent occupational heat exposure has been linked to cardiovascular and renal strain, including acute kidney injury; however, scalable and low-cost adaptation strategies remain limited. Farmer-managed natural regeneration (FMNR), a community-led agroforestry approach that promotes natural tree regrowth, can alter farm-level microclimates and generate a cooling effect, with potential to impact heat-related health outcomes. However, empirical evidence linking this ecosystem restoration to direct health outcomes is lacking. This study is a comprehensive impact evaluation examining the role of trees in croplands in reducing local heat exposure and heat-related physiological, cardiovascular, and renal strain among smallholder farmers in central Tanzania. Methods and Analysis We will conduct a quasi-experimental study involving approximately 500 adult farmers from villages implementing FMNR and matched comparison villages, who will be enrolled in a longitudinal panel design. Farm-level tree density assessments will be quantified using drone imagery and high-resolution satellite data on Trees Outside Forest. Farm-level microclimate measures such as temperature, relative humidity, and solar radiation will be measured using a combination of portable weather stations and remotely sensed data. Physiological outcomes will include continuous measurements of skin and core body temperature, heart rate, and blood pressure. Clinical outcomes will include pre- and post-work measures of hydration using urine samples, and acute kidney injury measured using serum creatinine. All data will be collected during peak farming seasons and repeated after 2 years. Multilevel mixed-effects regression models will estimate differences in outcomes between intervention and comparison sites, while difference-in-differences analyses will assess changes over time. Mediation analysis will evaluate whether changes in microclimate and tree cover explain observed health outcomes. Community Engagement A Community Advisory Board will collaborate throughout the study to ensure cultural appropriateness and guide context-specific procedures.</ns5:p>