Remote Sensing Applications within Ephemeral Channels: Quantifying Temporal and Spatial Flow Patterns within the Souss-Massa basin, Morocco

Intermittent and ephemeral channels are a critical component of the global hydrologic network, accounting for 30-50% of river length and discharge annually. Ephemeral channels are particularly relevant in dryland environments, where high rates of transmission loss facilitate aquifer recharge. Present literature highlights the relationship between antecedent streamflow conditions and focused recharge, as sediment moisture is tied to infiltration potential. Understanding shifts in surface flow has direct implications for estimations of aquifer recharge and storage. This can be impeded by limited stream gauging commonly associated with intermittent systems. Remotely sensed imagery provides an opportunity for improved temporal and spatial quantification of intermittent flow regimes and associated recharge. This is critical as increased global temperatures are expected to alter existing patterns of precipitation, with the potential to shift presently perennial streams to intermittent channels. This research is focused on identifying and quantifying shifts in patterns of flow intermittency, to improve estimations of aquifer recharge in a gauge-limited, ephemeral channel in central Morocco. The Souss-Massa Basin is highly agricultural and reliant on surface and groundwater pumping, experiencing significant droughts in the past decade. Discriminant function analysis (DFA) was applied to remotely sensed imagery (LANDSAT 5, 7, and 8) from 1984-2020 for water pixel classification, used to identify patterns of surface flow. Water identification within intermittent channels presents a challenge, as pixels may be shallow, turbid, mixed, or temporally variable. DFA allowed for optimization of pixel classification to site and sensor specifications, outperforming the commonly applied Mod. Normalized Difference Water Index (MNDWI). Preliminary analysis suggests the improved sensitivity of DFA, which demonstrates greater efficacy in identifying known water pixels than MNDWI. Findings further suggest shifting spatial variability and extent of surface flow across upstream, midstream, and downstream channel sections during the study period. With projected climate warming, developing tools to quantify the evolution of intermittent channels is necessary for sustainable water management.