Hydrological Modelling
Our hydrological research looks at the fate of water and dissolved chemicals within hydrological systems.
In particular we study the effects of land management and climate change on floods, drought, and diffuse pollution. We also combine models, process understanding and data to quantify and, where possible, reduce the uncertainty associated with predictions of future river levels and water quality.
General Projects - Current
Integration of environmental models and physical constraints with information theory and signal processing
- Current staff and students involved: Dr Bethanna Jackson
- Past staff and students involved: Dr Ilias Pechlivanidis (now at the Swedish Meteorological and Hydrological Institute)
- External collaborators: Dr Hoshin Gupta, The University of Arizona and Dr Hilary McMillan, National Institute for Water and Atmospheric Research
Quantitative theoretical models play a critical role in understanding and managing environmental resources and key related issues such as hazards and adaptation to climate change. However, current predictions are contaminated by significant data and model errors. To achieve the improvements demanded in modelling and forecasting, there is a need to overcome fundamental gaps in knowledge of earth system processes, and develop objective means of evaluating the value of various forms of data for both improving predictive capacity and advancing system understanding. VUW is working with colleagues at NIWA, at Arizona State University, and Swedish Meteorological & Hydrological Institute to extend information theory to respect physical constraints, enabling environmental scientists to (1) measure structure and extent of ‘memory’ within catchment systems; (2) reconcile process understanding and models with data; 3) quantify the value of data and hence better design monitoring networks.
We are developing entropy and time series analysis methods within the context of hydrological and earth systems science theory (mass/energy balance and other physical constraints). Entropy measures (e.g. Shannon entropy and mutual information) and principles (e.g. minimum entropy) provide formal quantifications of uncertainty and compressibility in data. Meanwhile, filters and frequency analysis are being used to examine dynamical structure and to distinguish between “coloured” (structured), and “white” (unpredictable) signal and noise components. Together, they provide complementary characterisations of the nature and extent of structure and hence predictability. We are using our results to characterise how inputs (e.g. precipitation, incident energy and chemicals) are transported and modified within hydrological systems (soils, aquifers, rivers, and the integrated catchments). We believe this work will enable us to identify and correct process representation and data insufficiencies resulting in improved forecast accuracy and precision and establish limits to predictability in catchments for varying degrees of data availability.
Publications
Pechlivanidis, I.G., Jackson, B.M., McMillan, H., & Gupta, H. (in press). Use of an entropy‐based metric in multiobjective calibration to improve model performance. Water Resources Research.
Pechlivanidis, I.G., Jackson, B.M., & McMillan, H. (2010). The use of entropy as a model diagnostic in rainfall-runoff modelling. In Hood S. (Technical Ed.) Proceedings of the 2010 International Congress on Environmental Modelling and Software. Manno, Switzerland: iEMS
Development of a flood forecasting framework, using data assimilation to improve the accuracy for streamflow forecasts
- Current staff and students involved: Dr Deborah Maxwell, and Dr Bethanna Jackson
Deborah is developing a flood forecasting framework to allow the continuous prediction of streamflow using real-time rainfall forecasts. The aim is to provide near-real time and sufficiently accurate streamflow predictions. To achieve this, she is combining a physically consistent conceptual model (although in the future she aims to be able to incorporate multiple model structures) and a data assimilation technique (a constrained Ensemble Kalman Filter) to improve the reliability of model output.
While the conceptual model is able to provide reasonable simulations, due to simplifications in the model structure, imperfect data driving the model and estimated parameter values, there are still likely to be trade-offs in some parts of the hydrograph. The constrained Ensemble Kalman Filter overcomes some of this uncertainty by updating model states to be more representative of antecedent conditions. This state-updating approach uses the information contained in real time or near-real time streamflow observations to correct for biases in modelled catchment storage, improving the reliability of forward predictions and allowing more robust decision making to be made.
Publications
Maxwell, D.H., Jackson, B.M., & McGregor, J. (submitted, under review). Constraining the Ensemble Kalman Filter for improved streamflow forecasting. Journal of Hydrology.
Recently Completed General Projects
Modelling impacts of climate change and land management on flood frequency and magnitude
- Current staff and students involved: Dr Bethanna Jackson, Dr Deborah Maxwell
- Past staff and students involved: Mr John Ballinger, Dr Ilias Pechivanidis
- External collaborators: Many - see listed co-authors in publication list
With changing climate, floods in New Zealand are likely to become more frequent and intense. This will result in increased risk to major infrastructure including failure of flood protection measures.
Hydrologists at Victoria University of Wellington are working with NIWA to investigate possible changes in river flows and extreme flood events due to climate change in New Zealand.
This project was funded by the Ministry for Primary Industries (formerly the Ministry of Agriculture and Forestry) Sustainable Land Management & Climate Change Research Programme.
We are working with the New Zealand Climate Change Research Institute to predict changes in flood frequency and magnitude given various climate scenarios, and producing associated flood inundation maps for the Hutt Valley, near Wellington.
Publications
McIntyre, N., Ballard, C., Bruen, M., Bulygina, N., Buytaert, W., Cluckie, I., Dunn, Ehret, U., Ewen, J., Gelfan, A., Hess, T., Hughes, D., Jackson, B.M., Kjeldsen, T.R., Merz, R., Park, J.-S., O'Connell, E., O'Donnell, G., Oudin, L., Todini, E., Wagener, T., & Wheater, H. (2014). Modelling the hydrological impacts of rural land use change. Hydrology Research, 45(6), 737-754.
Lawrence, J., Reisinger, A., Mullan, B., & Jackson, B.M. (2013). Exploring climate change uncertainties to support adaptive management of changing flood-risk. Environmental Science & Policy, 33,133–142.
Wheater, H.S., Ballard, C., Bulygina, N., McIntyre, N., & Jackson, B.M. (2012). Modelling Environmental Change: Quantification of Impacts of Land Use and Land Management Change on UK Flood Risk. In L. Wang & H. Garnier(Ed.), System Identification, Environmental Modelling, and Control System Design (Chapter 22, pp. 449-481). New York, USA: Springer.
McIntyre, N, Ballard, C., Bulygina, N., Frogbrook, Z., Cluckie, I., Dangerfield, S., …, Jackson, B., …, & Wheater, H (2012). The potential for reducing flood risk through changes to rural land management: outcomes from the Flood Risk Management Research Consortium. In Proceedings of the BHS Eleventh National Symposium, Hydrology for a changing world, Dundee 2012. UK: British Hydrological Society.
McMillan, H., Jackson, B.M., & Poyck, S. (2010). Flood risk under climate change: A framework for assessing the impacts of climate change on river flow and floods, using dynamically-downscaled climate scenarios (Technical Report). Wellington, New Zealand: NIWA for Ministry of Agriculture and Forestry, CHC2010-033.
Student Projects - Current
Groundwater age determination using a stochastic framework
Monique is developing a stochastic framework to infer groundwater age from environmental tracer data, such as tritium and SF6.
Groundwater age contains information about 1) when the water was recharged and 2) what mixing or flow processes it underwent. Therefore groundwater dating enables the understanding of the dynamic of a groundwater system, its sustainability and contamination potential. For example relatively young groundwater (<10 years) is more prone to anthropogenic contamination than relatively old groundwater (>100 years), due to only recently introduced contaminants, such as fertilizers, and less time for the contaminants to be removed by microbes.
The commonly accepted groundwater dating approach involves interpretation of environmental tracer observations, such as tritium, carbon 14 and SF6. Age is interpreted by considering their time dependent input to the groundwater system, a decay term (if applicable) and simplified lumped parameter models, which describe mixing processes of different aged waters in the aquifer or during sampling. The lumped parameter model, which contains the age information, is found by manual simulation of tracer concentrations as a function of the lumped parameter model until observed and simulated concentrations (visually) match.
This approach serves as a basis for Monique’s framework. She is extending it by including a Monte Carlo simulation and comparison of observed and simulated concentrations by means of an objective function. This allows assessment of the entire model space (in contrast to previous manual point simulation), objective comparison of modelled and simulated concentrations (in contrast to previous visual comparison) and assessment of the range of models, which equally well fit the observations under consideration of various input uncertainties, such as measurement error of observations in groundwater and recharge (in contrast to previous best fit approach, not considering input uncertainties).
The aim is to more conservatively constrain the age interpretation to its possible range by accounting for various uncertainties in model input. This will lead to a more reliable/secure groundwater management, such as assessment of sustainability or potential contamination. Subsequently the more adequately constrained age interpretation may help to establishing a concentration - time (age) relationship for various hydrochemistry parameters. These can then be used as new, inexpensive, more accessible and easier determinable tracers for groundwater age.
Monique is co-supervised by Dr Bethanna Jackson and Dr Uwe Morgenstern (GNS). Her project is part of and funded by the GNS Science project ‘SMART aquifer’ funded by the New Zealand Ministry of Science and Innovation.
Improving LUCI’s predictive ability to model nitrogen and phosphorus emissions to water
The aim of this project is to further develop LUCI’s ability to model nitrogen and phosphorus emissions to water associated with rural land management and primary production. New Zealand data, literature and understanding will be used to identify and quantify nutrient inputs to the model and develop new nutrient export algorithms for the model, which will consider a wider range of influencing variables than are currently considered.
Three levels of emissions to water tools using the improved algorithms will be developed to cater for varying levels of input data available and output data required. These will be applied to New Zealand catchments and the outputs explored to investigate spatially explicit mitigation solutions to the water quality problem.
Martha is supervised by Dr Bethanna Jackson and Dr Ants Roberts (Ravensdown) This project is part of wider LUCI research and development, and is supported by Ravensdown.
Irrigation modelling with the LUCI ecosystem services framework
Stuart Easton
Stuart is a Masters student who is developing an irrigation model to work within the LUCI ecosystem services framework.
Agriculture is highly reliant on irrigation inputs to farmland. However the quantity, timing and distribution of inputs are rarely monitored. Irrigation modelling therefore requires a predictive element that reflects on farm management strategies, as well as hydrological modelling of water and land interactions.
Stuart's work aims to help LUCI tools produce accurate nitrogen and phosphorus loss predictions, as well as provide irrigation efficiency measures and recommendations.
Stuart is co-supervised by Dr Bethanna Jackson and Dr Mairead de Roiste. He is supported by a Ravensdown funded project secured through Wellington UniVentures, developing a bespoke version of LUCI suitable for recommending detailed nutrient management at farm scale.
Recently Completed Student Projects
Groundwater age in the Wairarapa
Ryan Evison. Ryan was co-supervised by Dr Bethanna Jackson and Chris Daughney, GNS Science. View more here .
Mathematical modelling of solute transport in a heterogeneous aquifer
James Phillip Domisse. James was supervised by Dr Bethanna Jackson. View more here.