Toward long-range forecasts of waterbird...
URL: https://www.mdba.gov.au/sites/default/files/publications/Toward%20long-range%20forecasts%20of%20waterbird%20outcomes%20to%20enhance%20annual%20environmental%20water%20planning.pdf
Authors: David E Robertson, Jin Teng, Jorge Luis Pena-Arancibia, Andrew Schepen, Micha V Jackson, Heather M McGinness, Cate Ticehurst, Andrew Freebairn and Ashkan Shokri
This report demonstrates a proof-of-concept approach to forecasting wetland inundation at sub-seasonal to annual time scales and assesses its performance.
Models to predict the floodplain water volumes from observed inflows and antecedent conditions, and to predict the spatial extent of inundation from the floodplain water volume, were developed for both study regions. These models were then combined with inflow forecasts to forecast floodplain water volumes and maps of the probability of inundation.
The report highlights the next steps required to address limitations in the prototype forecasting system and to translate the system into an operational environment that would support real-time decision making.
Key findings / recommendations:
-
The multi-index method (MIM) imagery provided a fit-for-purpose estimate of flood inundation extents and gap-filled these images at the pixel level to produce near-complete monthly inundation timeseries for subsequent analysis.
-
Using established methods to generate inflow forecasts at monthly time steps, these forecasts were shown to skilfully predict inflows to lead times of up to 5-7 months throughout the year.
-
Forecasts of inflows into the Murray River were slightly more skilful than forecasts of inflows into the Macquarie River.
-
Analysis of retrospective floodplain water volume forecasts showed that these forecasts are skilful and provide a statistically reliable estimate of the forecast uncertainty.
-
A framework was established for forecasting waterbird nesting outcomes, by identifying the key cues for waterbird nesting that encompass indicators of wetland inundation and a range of other factors.
-
Through the investigation many technical limitations were identified that could be addressed by further research and development. In addition, the forecasting methods used in this study have only been used for retrospective analysis and for the forecasts to be used operational decision-making would require the deployment of real-time forecasting systems.
You can copy and paste the embed code into a CMS or blog software that supports raw HTML