Climate, vegetation cover and dust storm events in arid and semi-arid regions of Australia - relationships, models and predictions

Climate, vegetation cover and dust storm events in arid and semi-arid regions of Australia - relationships, models and predictions

Wind erosion is a land degradation process in arid, semi-arid and agricultural regions of Australia. The loss of soil as a result of this process affects human health, environment and the economy. Climate variables such as rainfall and temperature play a major role in wind erosion activity. In particular, the quantity and distribution of rainfall influences the growth of vegetation cover which protects the soil surface from erosion (both wind and water erosion). Hence, climate variability is of great concern due to the pressure on agricultural land to produce more food for a growing population and the subsequent pressure to grow crops on drier more marginal lands that are more susceptible to wind erosion.

This research investigates the historic relationship between climatic conditions and recorded dust storm events based on more than 16 decades of collated dust storm event data from a wide number of sources (e.g. personal experiences, diaries, book excerpts, newspaper clippings, journal articles, reports and others). The 587 dust storm event records have been collated into a Historical Dust Event Database (HDED). The HDED indicated an increased number of dust storm events occurred in the 1900s, 1940s, 1960s and 2000s. This is due to the close link of rainfall and temperature to the ENSO cycle which directly impacts on the vegetation cover, a key factor driving the frequency, intensity and spatial distribution of dust events.

Broad scale estimation of spatial changes in vegetation cover would be useful in a wide range of applications and is of particular interest and value in areas of environmental, ecological and land-use modelling. Currently, broadly applicable modelling methods or indices are not available to realistically estimate vegetation cover levels for periods before the early 1990s when satellite remote sensing first became readily available. This includes any historical or future forecasting periods. As wind erosion/dust events are strongly dependent on vegetation cover, to analyse past or future dust events a means of estimating broad scale cover across Australia is required.

The newly developed Climate Aridity Vegetation Index (CAVI) is a simple broad scale vegetation index across Australia, based on rainfall and temperature data. The CAVI is calculated using 12 months weighted rainfall and temperature data to produce vegetation cover maps without modelling individual vegetation type responses, seasonality and land-use. The CAVI produced particularly good estimation of vegetation cover during the Spring - Summer season but can over emphasise the relationship between rainfall, temperature and vegetation/green cover when increased rainfall occurs close to the month of interest. Nevertheless, the index produces good representative estimates and spatial maps of vegetation cover levels during the spring – summer seasons in Australia.

Wind erosion modelling occurs at a variety of spatial and temporal scales to determine the extent and severity of wind erosion across Australia. With the development of CAVI, historical and future wind erosion rates can be modelled, dust source areas can be estimated and identified, and the severity of these early dust storm events can be compared to modern events before land management changes were adopted. This has previously never been possible since reliable satellite derived photosynthetically active fractional vegetation cover (fPV) data is not available prior to February 2000. To test the validity of such models, CAVI estimates of vegetation cover have been tested as a surrogate for remote sensed fPV in the Computational Environmental Management System (CEMSYS) for two large scale dust storm events in September 2009 and October 2002. The CEMSYS estimated daily dust loads based on CAVI and fPV were compared in regards to the spatial patterns of the eroded areas and the dust load intensity of the modelled wind erosion days. The use of CAVI as a surrogate for fPV in September 2009 and October 2002 CEMSYS modelling results were encouraging. Similar spatial erosion characteristics were observed in the simulations but the dust concentration based the CAVI was on occasions lower than based on fPV data. The CAVI was also applied to model the historical dust storm periods in November 1965. The modelling results from the study indicates that there is potential for CAVI to be used as a surrogate for fPV and gives us for the first time some estimates of the extent and severity of historical dust storm events.