MACH-Flow

Motivation

‍Accurately predicting river discharge, in particular at ungauged locations, is crucial for water resource management as well as drought and flood monitoring. Traditional hydrological models are generally calibrated on data from a single river catchment. They perform quite well when forecasting at this particular location (satisfactory temporal prediction) but they often generalize poorly at other locations (unsatisfactory spatial prediction). The main goal of this project is to develop models which can predict in space as well as in time. To this end we make use of data from many river gauging stations within a river network and model them jointly.

Proposed Approach / Solution

‍We are currently developing two models in parallel. The first one is a recurrent neural network with long short-term memory (LSTM) cells, following the hydrological state-of-the-art architectures. Here, all gauging stations are considered independent and they share parameters (Figure 1). An extension we are currently developing is with a loss function based on a generalization of the continuous rank probability score (a kind of energy score). This in principle allows to build ensemble predictions which would quantify predictive uncertainty. The second model builds on first principles with an additive model with ridge-penalized cubic B-splines. This model includes engineered features for both short- and long-term effects of precipitation and temperature. We have an optional independent module which can incorporated to this additive model. This module routes water through a river network, the latter seen as direct graph. We have illustrated with a benchmark hydrological model that such spatio-temporal routing can only improve predictions, and this mainly for more downstream locations (Figure 2). A future research avenue is to combine this water routing module (graph-based convolutions) with a recurrent neural network.

Impact

‍First, our LSTM neural network is on par with state-of-the-art hydrological models for Switzerland, albeit with much less inputs. This allows us to reconstruct river discharge at a national scale back to the early 1960s, which is the earliest time for which precipitation and temperature data products are available. This reconstruction is important for the hydrological community as its low computation cost allows one to test various climate change scenarios. Second, the routing model, to be released as an independent module, is useful for the community as it can be easily added to any local prediction model to improve downstream predictions. In principle it can predict improve predictions even with discharge data only available at a single station. Finally, a version of our additive model will be soon added to the ensemble of models used operationally on the drought.ch drought monitoring platform maintained by our partners at WSL.