Using LSTM to monitor continuous discharge indirectly with electrical conductivity observations

Abstract

Due to EC's easy recordability and the existence of a strong correlation between EC and discharge in certain catchments, EC is a potential predictor of discharge. This potential has yet to be widely addressed. In this paper, we investigate the feasibility of using EC as a proxy for long-term discharge monitoring in a small karst catchment where EC always shows a negative correlation with the spring's discharge. Given their complex relationship, a special machine learning architecture, LSTM (Long Short Term Memory), was used to handle the mapping from EC to discharge. The results indicate, based on LSTM, that the spring's discharge can be predicted well with EC, particularly in storms when the dilution dominates the EC dynamic; however, the prediction may have relatively large uncertainties in the small or middle recharge events. A small number of discharge observations are sufficient to obtain a robust LSTM for the long-term discharge prediction from EC, indicating the practicality of recording EC in ungauged catchments for indirect discharge monitoring. Our study also highlights that the random or fixed-interval discharge measurement strategy, which covers various climate conditions, is more informative for LSTM to give robust predictions. While our study is implemented in a karst catchment, the method is also suitable for non-karst catchments where there is a strong correlation between EC and discharge.

Transformer Versus LSTM: A Comparison of Deep Learning Models for Karst Spring Discharge Forecasting

“…EC has generally such a strong correlation with discharge, that it can be used as proxy for discharge, as shown for example, by Chang et al. (2022) for a small karst catchment in China (Chang et al., 2022). In our study it improved four of the six models in the noQ scenario when used as input feature, particularly for the Aubach spring which has a short response time.…”

“…The physicochemical variables are well established for karst systems. They were found to correlate moderately with discharge at LKAS2 (correlation coefficients of −0.70 for EC, 0.67 for turbidity, and 0.66 for SAC, p ‐value < 0.05, Reischer et al., 2008), but also in other karst catchments (e.g., Chang et al., 2022). Meteorological variables were available from two stations in the catchment area at altitudes of 1,350 and 1,520 m a.s.l.…”