Willem Vervoort is Associate Professor in Hydrology and Catchment Management, and Director of the ARC Training Centre in Data Analytics for Resources and Environments (DARE), both at the University of Sydney. His discipline expertise is in developing and improving models to advance sustainable water management for better decision making. Willem’s technical expertise focuses on improving simulation models of catchments at different scales. This includes quantifying uncertainty in predictions, integrating big data and remote sensing in models, and understanding how the real world is simplified in models. Willem also has strong industry collaborations in research, consultancy projects, and international collaborations in Europe, Indonesia, Uruguay, and India.
Abstract
Catchment modelling and data science for sustainable catchment management
Developing sustainable management solutions at the landscape level requires balancing different activities which overall results in the lowest impact. Given the additional overlying impact of climate change and uncertainty in observations, it can often be difficult to see how individual management actions impact landscape level outcomes.
Over the last 10 years, we have developed several approaches to these problems. A key element in these approaches has been to use modelling and statistical analysis to separate climate signals from management actions. Combining conceptual modelling with residual trend analysis, the influence of climate change, reforestation and increasing water licences can be separated. This builds on earlier work, where we demonstrated that the “amplification” of a drying in rainfall trend in runoff is a minor component in forested catchment. This might suggest some sort of adaptation in the vegetation that counteracts observed amplification in more diverse landuse catchments.
Current work involves a re-analysis of a large data base of catchment studies focussed on reforestation and deforestation. Previous work finds a strong relationship between forestation/de-forestation and streamflow but different confounding factors on the impact of forestation. Generalised additive models were used to run flexible regressions including multiple variables on the dataset. The results indicate that, changes in forest cover still cause changes in streamflow, however this change is different for deforestation and reforestation, with deforestation indicating larger changes in streamflow. The area of the catchment affected the results, but this is potentially caused by the wide variety in reported results from small scale paired catchment studies (Figure 1). These smaller studies dominate the database with 42% of the data < 1 km2 and 65% of the data < 10 km2. As a result, the paired catchment study assessment technique increased the change in flow by 135% in the model.
Ongoing work focuses on the effect of increased streamflow on floodplain vegetation and analysing whether the evapotranspiration signal has changed over time in a large complex catchment.
Overall, this work indicates that separating management actions from climate variability in large complex catchments is difficult ad requires careful consideration before conclusions are drawn.