CEMSYS modelled wind erosion

Leys, John F. and Butler, Harry J. ORCID: https://orcid.org/0000-0002-1224-2774 and Yang, Xihua and Heidenreich, Stephen (2010) CEMSYS modelled wind erosion. Project Report. NSW Department of Environment, Climate Change and Water , Sydney South, Australia. [Report]

[img] Text (Published Version)

Download (1MB)


[Executive Summary]:
The Leys report on wind and water erosion (Leys et al. 2009b) recommended that wind erosion modelling be undertaken to assist in reporting the extent and severity of wind erosion across Australia. The modelling could then be used by the Australian Government, states and Natural Resource Management (NRM) bodies for resource condition reporting. The same products could be used to assist in identifying areas for Caring for our Country (C4oC) investments.

Modelled monthly and annual wind erosion maps of Australia at 50-km resolution for the period July 2006 to June 2008 were compiled using the Computational Environmental Management System model (CEMSYS). CEMSYS comprises an atmospheric model, a land surface model, a wind erosion model, a transport and deposition model and a land surface database. It uses analysis data from the National Centre for Environmental Prediction, USA (NCEP) to calculate the atmospheric properties like wind fields, rainfall, radiation and clouds. Geographic Information Systems (GIS) are used to describe soils and vegetation data and monthly satellite data is used to calculate ground cover levels. In this study, the severity of wind erosion is described by the horizontal soil flux (TQ mg/m/s) output from CEMSYS. TQ is representative of the average amount of soil that is moved by wind within the pixel each month.

To aid with reporting, a modified map of NRM regions and subregions was developed by an expert panel to report wind erosion status. The severity of erosion, expressed as five erosion classes (very low to very high), was then calculated for each subregion, region, state and the continent for 24 months.

In the 2007–08 dust-year 11% of Australia was in the high and very high erosion classes. This compares to 9% of Australia in the 2006–07 dust-year; however, this 2% yearly difference is not statistically significant. NRM regions with the largest areas of erosion (high and very high classes) tend to be focused in arid and semi-arid rangelands of south-western Queensland, western NSW, north-central and north-eastern South Australia and western Western Australia. The semi-arid agricultural lands of eastern West Australia also had areas of high and very high class erosion. Notably, the non-agricultural lands of western South Australia, northern Northern Territory and eastern Western Australia all have low erosion levels.

The NRM regions, by state, with the highest amounts of soil moved (TQ) were:
 Desert Channels, South West, Border Rivers Maranoa–Balonne, and Condamine in Queensland
 Western, Border Rivers–Gwydir, Namoi, Lachlan, Murrumbidgee, Murray and Lower Murray–Darling regions in NSW
 the Arid Lands and Northern and Yorke regions of South Australia
 the Pastoral and Non-Pastoral regions of the Northern Territory, and
 the Rangeland regions of the Goldfields Nullarbor, Gascoyne Murchison and the mixed farming Northern Agricultural and Avon regions of Western Australia.

The CEMSYS map outputs produced in this study offer a greater temporal (monthly) and spatial resolution (50 km) and better statistical descriptions than the previous measure of wind erosion, the Dust Storm Index (DSI). DSI maps are derived from Bureau of Meteorology observer data from 110 sites across Australia at annual time steps. Maps are then created by interpolating between the 110 sites. Despite the temporal and spatial limitations of the DSI, it still remains a very valuable cross validation for CEMSYS and it has the major advantage of a longer time series (1960 to the present).

With the exception of the annual DSI data, there is a lack of wind erosion data to test the model outputs against at the 50 and 10-km scales. Roadside survey data from NSW is one- hectare scale data and was used to determine the five erosion severity classes and therefore could not be used to test the model. Future testing is planned after the compilation of DustWatch data and roadside survey data from other states (South Australia and Western Australia).

Model outputs appear to be most reliable in the south-eastern Australian rangelands. Outputs for the savannas and grasslands of northern Australia seem less reliable due to the lack of accuracy of the satellite ground cover product’s ability to detect dead or senescing vegetation. This error in cover estimates results in an over-prediction of erosion in northern Australia. Under-prediction of erosion in the winter cropping areas of southern Australia is noted and possibly relates to the simplification of soil types used in the model. The model’s performance will be improved by using better ground cover and soils data and projects are underway to address these data issues.

The issue of scaling needs to be appreciated and this project shows the problems of comparing data measured at different scales (10 and 50 km) and the loss of precision at 50 km. Larger pixels involve greater averaging of underlying information and as such the likelihood of reporting high and low values is reduced. Therefore, the chance of locating areas with high levels of wind erosion is reduced. In this study, the scalded river margins along the Edward and Wakool Rivers in the western end of the Murray catchment management area were classified ‘High’ and ‘Very High’ at 10-km resolution and ‘Low’ and ‘Moderate’ with the 50-km resolution. With the 50-km data, this priority area would not have been identified.

This study concludes with an implementation plan that proposes improvements to the modelled wind erosion data through the provision of a longer time series of maps (2000 to the present) and increasing the resolution of the modelling from 50 to 10 km.

Statistics for USQ ePrint 8781
Statistics for this ePrint Item
Item Type: Report (Project Report)
Item Status: Live Archive
Additional Information: Copyright © 2010 Commonwealth of Australia. PRN 0808 – 1220.
Faculty/School / Institute/Centre: Historic - Faculty of Sciences - Department of Maths and Computing (Up to 30 Jun 2013)
Faculty/School / Institute/Centre: Historic - Faculty of Sciences - Department of Maths and Computing (Up to 30 Jun 2013)
Date Deposited: 07 Oct 2010 02:30
Last Modified: 17 Nov 2021 03:57
Uncontrolled Keywords: wind erosion modelling; Computational Environmental Management System model; CEMSYS
Fields of Research (2008): 05 Environmental Sciences > 0502 Environmental Science and Management > 050205 Environmental Management
05 Environmental Sciences > 0502 Environmental Science and Management > 050206 Environmental Monitoring
Fields of Research (2020): 41 ENVIRONMENTAL SCIENCES > 4104 Environmental management > 410404 Environmental management
41 ENVIRONMENTAL SCIENCES > 4104 Environmental management > 410404 Environmental management
Socio-Economic Objectives (2008): D Environment > 96 Environment > 9606 Environmental and Natural Resource Evaluation > 960609 Sustainability Indicators
D Environment > 96 Environment > 9606 Environmental and Natural Resource Evaluation > 960699 Environmental and Natural Resource Evaluation not elsewhere classified
URI: http://eprints.usq.edu.au/id/eprint/8781

Actions (login required)

View Item Archive Repository Staff Only