A semi-automated remote-sensing method exploring the dewatering risk of reservoir seasonal drawdowns on Lake Trout (Salvelinus namaycush) spawning site

Abstract

Seasonal water-level fluctuations associated with hydroelectric reservoir management can generate littoral disturbances, with consequences for the sustainability of Lake Trout ( Salvelinus namaycush ) populations across North America. Assessing the dewatering risk of critical habitat such as spawning sites at fine spatial scales requires detailed nearshore water-level data. However, such data remains largely unavailable in many reservoirs, as their acquisition through conventional field methods is both costly and labour-intensive. This paper presents a semi-automated remote sensing approach for mapping water-level-specific shoreline contours in waterbodies lacking shoreline bathymetric data. The Direct Difference Water Index (DDWI) was identified as the most accurate index and was applied on PlanetScope multispectral satellite imagery to delineate the land-water boundary across multiple water surface elevations in a lacustrine reservoir with heterogenous shoreline substrate subject to seasonal drawdowns. By coupling PlanetScope derived shorelines with hydrometric records, water-level-specific shoreline contours were used to estimate potential littoral exposure. The methodology was applied at the Manouane Reservoir (Québec, Canada), both at fine scale over a putative Lake Trout spawning site and at a broader scale over the main reservoir basin. This approach is only applicable where minimum water levels coincide with cloud-free and ice-free high-resolution multispectral imagery but allows satellite imagery to efficiently capture shorelines at their lowest elevations. Results demonstrate the potential of this approach to provide approximate nearshore exposure mapping in areas where such information is often unavailable. This offers a rapid, remote and cost-effective tool for screening potential exposure risk of reservoir drawdowns on littoral-dependent fish populations.