Feature importance measures to dissect the role of sub-basins in shaping the catchment hydrological response: a proof of concept

Abstract

Understanding the response of a catchment is a crucial problem in hydrology, with a variety of practical and theoretical implications. Dissecting the role of sub-basins is helpful both for advancing current knowledge of physical processes and for improving the implementation of simulation or forecast models. In this context, recent advancements in sensitivity analysis tools could be worthwhile for bringing out hidden dynamics otherwise not easy to distinguish in complex data driven investigations. In the present work seven feature importance measures are described and tested in a specific and simplified proof of concept case study. In practice, simulated runoff time series are generated for a watershed and its inner 15 sub-basins. A machine learning tool is calibrated using the sub-basins time series for forecasting the watershed runoff. Importance measures are applied on such synthetic hydrological scenario with the aim to investigate the role of each sub-basin in shaping the overall catchment response. This proof of concept offers a simplified representation of the complex dynamics of catchment response. The interesting result is that the discharge at the catchment outlet depends mainly on 3 sub-basins that are consistently identified by alternative sensitivity measures. The proposed approach can be extended to real applications, providing useful insights on the role of each sub-basin also analyzing more complex scenarios.

Sensitivity and feature importance of climate factors and evaluation of different machine learning models for predicting fire hotspots in Kalimantan, Indonesia

“…This is different from the sensitivity of each feature where all features have sensitivity values that are almost close to one another. This shows that based on the sensitivity value, all features have an impact on hotspots because they have a correlation [10]. Thus, a small feature importance value does not mean that the feature has no effect at all on the hotspots in Kalimantan.…”

“…In a random forest, every tree receives independent predictions after being trained on a a random selection of features. By averaging the decision trees' projections, the response variable's final estimation is determined [10]. Figure 2(a) displays an example of the random forest model.…”

“…We train a decision tree at each step using the residuals from the preceding tree series. The additive model described by each tree's contribution is used to build the resulting ensemble model [10]. An illustration of the gradient boosting model is shown in Figure 2(b).…”

“…Thus, the effect or influence of the predictors mentioned above is not seen in more detail on forest fires. Analysis of sensitivity and feature importance has been carried out [10] to examine how sub-basins affect the hydrological response of catchments.…”

Flood hazard mapping using a GIS-based morphometric analysis approach in arid regions, a case study in the Red Sea Region, Egypt

“…Estimating morphometric parameters for a basin holds a significant importance as these parameters govern its hydrological response and provide valuable insights into its hydrogeological characteristics. Therefore, studying morphometric analysis and understanding a basin's response to heavy rainfalls, storms, and floods are vital for effective flood risk analysis (Diakakis 2011; Cappelli et al 2023). In data-scarce regions, morphometric properties of basins are commonly utilized to assess flood risks (Elsadek et al 2019a).…”