1. Introduction

Irrigation expansion raises schistosomiasis transmission by extending the habitat of the intermediate snail host and increasing human water contact [1], [2]. Where control resources are scarce, knowing the fine-scale geography of risk allows interventions to be targeted rather than blanket-applied. We mapped infection around the perennial canal network of the Awash basin and quantified how risk falls with distance from water.

2. Study area and data

We linked georeferenced parasitological surveys of 2,860 schoolchildren to a digitised map of perennial and seasonal canals. For each child we computed the straight-line distance to the nearest perennial canal and to the nearest canal junction, along with household water source and latrine access.

3. Statistical methods

We modelled infection with a multilevel logistic regression, with children nested in villages:

$$\operatorname{logit}\big(\Pr(y_{ij}=1)\big) = \beta_0 + \beta_1 d_{ij} + \beta_2 j_{ij} + \gamma^{\top} X_{ij} + u_j,$$

where $d_{ij}$ is distance to the nearest perennial canal, $j_{ij}$ is distance to the nearest junction, $X_{ij}$ are individual covariates, and $u_j \sim \mathcal{N}(0, \sigma^2)$ is a village random effect. Odds ratios are reported as $e^{\beta}$.

4. Results

Infection clustered sharply near water: prevalence exceeded 30 percent within 500 m of a perennial canal and fell to background beyond 1 km (Table 1). Proximity to a canal junction was the strongest single predictor, with an adjusted odds ratio of 2.7 for children living within 200 m of a junction.

5. Discussion and conclusion

Transmission is concentrated within a few hundred metres of perennial canals and around junctions where water contact is greatest. This geography argues for focal snail control and safe-water-contact points at identified hotspots rather than uniform coverage, complementing mass drug administration [2]. The cross-sectional design cannot establish direction, and distance is a proxy for exposure; longitudinal water-contact studies would refine the risk surface.