1 Centre for Resource and Environmental Studies, The Australian National University
2 Integrated Catchment Assessment and Management Centre, The Australian National University
3 Spatial Solutions
4 Department of Forestry, The Australian National University
Phone: (02) 6125 8129, Fax: (02) 6125 0757
Managing water resources to ensure environmental values are maintained, whilst allowing for continued economic development is a major challenge facing many areas including the Australian Capital Territory (ACT). This paper reports on a GIS based investigation of the implications of land use change on ACT water use. The paper describes a suite of tools that are collectively termed PLUCA (Platform for Land use Change Assessment).
Areas with the potential for land use change were identified through land capability assessment and by investigation of the suitability of land for development of alternate industries. Spatial data including slope, aspect, a wetness index, climatic surfaces, geology and consideration of the minimum viable scale of industry were analysed in the study.
A coarse land use class – water use relationship estimated for the ACT was used to determine the maximum potential water use resulting from land use change. Three scenarios, based on different levels of land use change were constructed to simulate high, medium and low levels of potential landuse change in the ACT. The estimated reduction in streamflow for the maximum development scenario, was around 6.8% of the average annual runoff from the ACT. This scenario represented modification of only 3.9% of the total land area.
This study demonstrates the potential for the use of GIS in the optimisation of landuse from biophysical characteristics. The implications of such changes should they occur were calculated through investigation of the annual average reduction in streamflow. The study demonstrates the use of GIS techniques in quantifying interactions at appropriate scales for decision making. The development of improved decision support tools is also outlined.
The ACT’s Water Resources Management Plan, established under the Water Resources Act 1998, is a key component of the Territory’s water management strategy. The purpose of the plan is to provide the ACT government with a decision-making framework and strategic direction for the long-term management of its water resources.
The Plan aims to:
- ensure that the use and management of the ACT’s water resources sustain the physical, economic and social well being of the people of the Territory while protecting ecosystems;
- protect waterways and aquifers from damage and reverse damage that has already occurred; and
- ensure water resources are able to meet the reasonably foreseeable needs of future generations.
Management of the ACT’s water is made more difficult by the Murray-Darling Basin Commission’s (MDBC) cap on water extraction – and the ACT’s part in this – which was being negotiated in mid-1999. The cap and associated water reforms includes the requirement for States and Territories to set up a comprehensive water allocation system, including the allocation of water for the environment, which encourages the highest value sustainable use of water.
In the context of the final aim of the Water Resources Management Plan, the question of possible water demands from intensive, high-value agricultural and horticultural industries was raised. Viticulture had expanded considerably in the region and had the potential to grow further. Would implementation of the MDBC’s cap impact on the prospects for growth of water-dependent industries? What were the bounds of possible future non-urban water demand in the Territory? Without the luxury of time to conduct a detailed integrated study of future possible water demand, was there a ‘first cut’ method that could give an idea of the possible maximum water demand from intensive agriculture?
The Integrated Catchment Assessment and Management Centre of The Australian National University was contracted by the ACT government to develop and apply an assessment methodology which could give bounds on maximum possible future water use. This paper reports on that investigation. The project was a GIS based desktop study that aimed to determine the maximum potential water use across the ACT given complete development and intensification of agriculture over all available areas.
As background to the study a short description of physical environment of the ACT is presented. Details of the methods of the study, development of land use change scenarios and results follow. The project outcomes are discussed and a description of new and improved decision support tools for land planning under development is then made.
The Australian Capital Territory is located in the Southern Tablelands of South Eastern Australia. The Territory is wholly located within the Murrumbidgee River catchment – part of the Murray-Darling Basin. Forestry and agricultural areas presently cover 37% of the total area of the ACT. These areas are available for agricultural intensification.
The average annual water available from ACT’s controlled catchments totals 465 GL (Environment ACT 1999). Of this, 272 GL is designated for environmental flow, leaving 193 GL available for consumptive use. Existing water use totals about 65 GL and provision is made for future allocations of around 1.9 GL due to mounting agricultural demand and a further 6.5 GL for additional water supply over the next 10 years, leaving 120 GL unallocated.
The ACT area has a climate with cool to mild winters with warm summers. Average annual precipitation ranges from 605mm to in excess of 1000mm in the mountainous areas in the southern portion of the territory. Rainfall is generally uniformly distributed throughout the year.
Parent material and topography are the major determinants of soil distribution in the Canberra region (Sleeman et al. 1979). The geology of the area is diverse and generally Palaeozoic in age (Strusz 1979). It includes a range of felsic, igneous and volcanic rocks and slightly metamorphosed sediments. The area in the north east of the territory is largely low lying, undulating country with an average elevation of 600m. The southern and central western regions are more dissected and elevated. Generally soils display a texture contrast and range from shallow and stony soils with red subsoils on crests and ridges, to deeper yellow colluvial and alluvial duplex soils that are often sodic in the A2 and B1 horizons in drainage lines. Soils on lower and mid-slope sections have the most potential for agricultural development.
Three base data sets were used for the spatial analysis of the study: a 40m digital elevation model (DEM), digital geology and land use coverages. This data was supplied by Environment ACT. Annual average runoff figures, used in the water modelling, were taken from the ACT’s Water Resources Management Plan (Environment ACT 1999).
Figure 1 provides a schematic diagram of the analysis undertaken in this study. All spatial data analysis, excluding generation of BIOCLIM surfaces, was performed using ArcInfo.
Figure 1 Schematic diagram of data sets and analysis undertaken to estimate the reduction in ACT streamflow consequent on land use intensification scenarios.
Digital elevation data was analysed to derive a number of derivative layers. BIOCLIM surfaces of rainfall and minimum annual temperature were computed using the BIOCLIM program. BIOCLIM is a bioclimatic prediction system that uses parameters derived from mean monthly climate estimates, to approximate energy and water balances at a given location (Nix 1986). Two outputs from BIOCLIM were used in the study, a mean annual rainfall surface and a surface of the mean minimum temperature of coldest week.
A topographic wetness index (TWI) adapted from Moore et al. (1993), Equation 1, was calculated for the study area. This index is related to the wetness of a site and thus is also correlated with a number of other physical processes associated with plant growth, species occurrence and soil characteristics (Moore et al. 1993).
where is the Topographic Wetness Index; is the specific area of the catchment and is slope in radians.
Slope and aspect data layers were also produced from the DEM.
Land Use Intensification
Analysis of land tenure, land capability assessment and the development of a viticulture suitability index were used to identify land available for intensification of use. Analysis of tenure simply excluded urban and conservation areas from further analysis. Only existing forestry and rural areas of the ACT were considered in this study. Land capability assessment aimed to identify the potential of land for intensification of agricultural activity. Viticultural suitability analysis identified areas within the ACT with a high potential for viticultural development.
Land Capability Assessment
The land capability classification scheme of the NSW Department of Land and Water Conservation was adopted in this project, see Emery (1988). Land capability classes were derived from characteristics of parent material and topography – the major determinants of soil characteristics in the region (Sleeman et al. 1979). Slope classes were produced from the slope surface in accordance with the Emery classification scheme. These slope classes were combined with the geology coverage to create a layer of unique combinations of slope and geology – 33 combinations in total. Using expert knowledge, each of these unique combinations of slope class and geological type were ascribed to a land capability class to reflect their potential for land use intensification. A map of the land capability classification is shown in Figure 2. The map extends over the whole of the forestry and rural areas of the ACT.
The land capability classes were re-categorised into two categories of water use and one category where it was assumed that no potential for a change in land use exists. Land with potential for intensive agricultural development was derived directly from the most productive class of the land capability classification – Class 1. This class is characterised by slopes less than 2% with geological substrate that develops productive soils. By aggregating classes 2, 3 and 4 of the land capability classification, a second class, suitable for horticultural development was derived. This class comprised land that was excluded from intensive agricultural development, land with a slope of up to 15%, with geological substrate that develops more productive soils and land with slope no greater than 5%, with less productive soils. The remaining classes were deemed to have no potential to support land use with higher water use than at present. However, rural areas in these poorer classes were included in the calculation of reduction in streamflow caused by construction of farm dams.
Figure 2 Land capability classification
Site Selection for Viticulture
There is little reliable information available in the literature as to optimum and limiting conditions for viticulture under ACT conditions. It is acknowledged that the occurrence of frost, soil characteristics (including fertility and drainage) and water supply are most likely to influence production. Our approach was to develop a multi-criteria index of suitability for grape production to identify areas that have the potential to support viticulture. Slope, temperature, aspect, geology, TWI and rainfall were input variables to the index.
If one variable absolutely constrained production, for example, slopes greater than 15%, the area was excluded from further analysis. The scores for each variable were then weighted to reflect the relative importance that each variable plays in determining the success of viticultural production.
The distribution of areas prone to heavy frost in the landscape is difficult to predict at an appropriate scale. Our approach was to use minimum temperature data produced as part of the BIOCLIM program as a surrogate to identify frost prone areas at the broad scale. To supplement this analysis at a smaller scale, a constraint was placed on areas with a high TWI value. The TWI – a measure of drainage area and wetness, was used to constrain areas where cold air is likely to pool and vineyards most likely suffer damage from frost. Aspect was also considered in the frost analysis with northerly aspects weighted more favourably in the multi criteria evaluation.
Slope was considered to be the most important factor in the multi-criteria analysis due to its influence on soil characteristics – particularly drainage, and thus received the greatest weighting. Temperature and aspect were weighted equally, these characteristics were assigned a weighting around three times that of rainfall to correspond with their influence on potential production and probability of frost occurrence. Geology, as a determinant of soil characteristics, and TWI were ascribed an equal weighting, approximately double that ascribed to rainfall. The level of mean annual rainfall was given the lowest weighting, as irrigation is likely to be required to maintain production throughout the region.
A spatial representation of the viticulture suitability index was generated to include all areas where no variables absolutely constrain productivity. The remaining areas were then divided into two classes according to their score. Only the highest class was considered to have viticultural development potential. This reflects the reality that low scoring areas would be economically marginal. To further identify land with practical potential for establishment of vineyards, areas of less than 20 ha in size were eliminated, because small non-contiguous blocks were considered unmanageable for landholders. Thus only areas without constraints, with a high multi criteria score and a contiguous area of greater than 20 ha are included as viticulture areas in the water use modelling.
Water Use Modelling
The results of the land use intensification were used as an input to water use modelling. Viticulture took precedence in allocation of land use in the water use modelling, this was followed by intensive agriculture and then by horticulture – ordering firstly by water use then by anticipated economic value. Land use type and the use of on-farm water storage areas are thought to be the two most significant factors to affect water use and hence streamflow volume.
Due to the number of variables known to influence runoff and unknown infiltration and groundwater variables the use of streamflow data to establish a relationship between landuse and runoff for the estimation of evapotranspiration for each land used type was not possible in the time available. Information needed to account for this variability was generally unavailable or at an inappropriate scale.
Runoff was calculated for the study area using Vertessy et al. (1999) equations for grasslands and pine forests for rural and forestry land use types respectively. The input for these equations was annual precipitation (mm), which was determined from the BIOCLIM surfaces. An average figure for the whole of the ACT was used. This runoff figure was then used to estimate streamflow reduction due to farm dams using the relationships of Schreider et al. (1999).
Given the area of land that would be occupied under each scenario it was possible, using figures assumed to be the average annual water use of each land use type (see Table 1) to calculate landscape water use figures and the impact that these increased water uses might have on streamflow.
Table 1 Assumed water use figures for land use categories.
Water use ML/ha/yr
The three land use scenarios were combined with the water use figures in Table 1 to produce three scenarios for water use.
Scenario 1 – Water Use Under Current Land Use Regime
Using the land use layer for only the area occupied by both rural and a forestry land use type, the current water use for the area of interest was calculated.
Scenario 2 –Water Use Under Medium Land Use Change
By taking the predicted potential distribution of intensified land use and then halving the total areas of each land use type an estimate of medium potential land use change was calculated. These land use areas were then multiplied by their water use figures to provide an estimate of total water use.
Also added to the total water use estimate of the medium land use change scenario was the reduction in streamflow caused by construction of farm dams. Schreider et al. (1999) found that a farm dam density of 3.78 dams km-2 caused a 24.4% reduction in streamflow. Streamflow has been assumed to equal runoff. A percentage of 24.4% reduction in streamflow was applied to the runoff figure calculated using Vertessy et al. (1999) runoff equations and changes in water use which influence runoff.
Scenario 3 – Maximum Water Use Under High Land Use Change
Water use was estimated by modelling for a land use regime where all land with potential to change was developed for that purpose. The land areas occupied by each land use type were multiplied by their respective water use figures to produce an overall maximum potential land use change impact on water use in the ACT.
Under this scenario farm dam influences were included by using the Schreider et al. (1999) 32.3% percent reduction in streamflow of 6.07 dams km-2. This is equivalent to the farm dam density in the nearby Yass River catchment in 1998. This same percentage was applied to the runoff figure calculated using Vertessy et al. (1999) equations.
Water Use Modelling Results
The results of the water use modelling are presented in Table 2. The table presents calculation of the reduction in streamflow for each water use scenario.
Current Water Use
Annual average current water use is estimated to be approximately 35131 ML/yr for the agricultural and forestry land uses of the study area (a total of 18237 ha). Using Vertessy et al. (1999) calculations it is estimated that from the forestry areas no runoff will occur on average, while within the rural landscape, it is estimated that the runoff rate would be 1.5 ML/ha/yr (see Table 2).
Medium Water Use
By developing half of the land with potential to support a higher water use activity, that is 4616 ha or almost 2% of the territory, it is estimated that this could reduce total runoff from the ACT by approximately 3.5%: an extra 14812 ML/yr of water use (see Table 2).
High Water Use
By devoting all land that has the potential to support land use change to a higher water use activity, that is 9232 ha or 3.9% of the total area of the ACT, it has been estimated that a reduction in total runoff from the ACT could be as high as 28436 ML/yr, representing 6.8% of current runoff from the whole of the ACT (see Table 2).
Table 2 Water use by land use types under different land use change scenarios.
Land use type
Current Water Use (ML)
Medium Water Use (ML)
High Water Use (ML)
% of current use
Flow reduction as % of ACT Runoff
The results presented in this paper provide a useful guide for water quality planning in the ACT and more broadly for the Murray-Darling Basin. Of interest is that the estimated reduction in streamflow consequent with maximum development represents a significant proportion of the runoff of the ACT when considered in the context of the small area of land actually changed. This shows that studies of this nature are important for natural resource management to indicate general management directions and future investigations.
The project combined GIS analysis of land suitability and siting of land use change with basic water use modelling to estimate potential reduction in streamflow volumes. The GIS component of the project was based on widely accepted, relatively simple techniques of spatial analysis. The water use modelling was similarly simple, predominantly because of the time constraints imposed on the project. This section discusses the strengths and weakness of the project and makes recommendations on a second generation of tools for hydrologic assessment of land use change.
Figure 3 Schematic of potential improvements to the methodology of assessment of hydrologic land use change impacts.
A major output of the project was land capability mapping extending over the forestry and rural areas of the ACT. A usual criticism of land capability mapping is that existing land uses are considered when land capability classes are assigned. This was not the case in this project, only geology and slope were considered in the allocation of land capability classes. Inclusion of field-based or spatial soil data could further improve the land capability assessment process of this nature in future projects.
A second major output of the project was the calculation of a grid-based viticulture suitability index extending over the whole of the ACT. The index, based on multi-criteria evaluation, considered many factors including topography, rainfall and temperature. Of particular note was that surfaces generated by the BIOCLIM program – an effective spatial means of incorporating climate information, were used in the analysis. To improve the viticulture suitability index an improved method of frost prediction focused at the site scale is required. As in the case of the land capability mapping, inclusion of field-based or spatial soil data particularly soil drainage measures, could also improve viticultural suitability and similar evaluations. Finally, experimental data on the optimum and limiting conditions of viticultural establishment, based on experimentally derived data would also prove valuable for future work.
Given the rapid development of the viticulture industry within the ACT region, the project focused on deriving a viticulture suitability index. It would be possible in future projects to apply the approach taken here to other crops if the appropriate characteristics were known.
A major limitation of the project was that all water use modelling was undertaken using average annual streamflow totals. No examination of the temporal complexity of the underlying data was made in the study. This limitation is especially apparent when the aims of the water resources management plan relating to environmental flows are considered. Demand for water is very often negatively correlated with supply with peak demands occurring during low flow conditions. Thus, there is a need to consider the temporal aspects. One approach to improve the accuracy of this type of study could be through the comparison of flow indices for land use change scenarios. Flow indices are statistical measures of environmental and ecological characteristics of the flow regime.
A greater capacity to predict temporal changes may be through the use of rainfall-runoff modelling. In future applications the IHACRES rainfall-runoff model, which is widely applicable for hydrologic analysis and management across a range of catchment scales may be used. IHACRES is a hybrid empirical-lumped parameter process model that can be used to estimate streamflow from inputs of rainfall and temperature data. It has been tested comprehensively throughout the world and reported widely in the open literature, see (Jakeman et al. 1990; Evans et al. 1997; Croke et al. in preparation). The model has the potential to simulate changes in vegetation and land use lumped at the catchment scale. Continued work on improving the land use change capacity of the model is ongoing and may include disaggregation of catchments into relatively homogenous hydrologic response units to improve spatial results.
The analysis presented here did not investigate the economic or social implications of conversion to new land uses. An economic modelling capacity could potentially allow the investigation of the feasibility of shifting production to new land uses by consideration of structural adjustment costs. Economic analysis could also examine the scope for markets to adsorb increased production and estimate expected returns. This analysis would strengthen the development of land use change scenarios hence changes in water use. An example application of the use of IHACRES modelling together with economic modelling tools can be found in the work of Gilmour (2000) and Gilmour et al. (2001).
Increased interaction with industries, for example the wine industry could improve the estimates of water use and result in more accurate streamflow reduction totals. Closer interaction with industry would be especially important were economic analysis of the type described above be incorporated into future projects.
The project attracted considerable media interest with newspaper, radio and television coverage. The interest in the project stemmed primarily from potential conflicts between upstream and downstream use of water. The uncomplicated nature of the research also made the results more widely understandable.
This project was an attempt to integrate physical and economic information commonly used in catchment modelling through the use of a GIS. There is a need to improve knowledge of linkages between catchment behaviour, hydrology and economic evaluation. This is a weakness of the work presented here that should be improved in future work of this nature. Hydrology is inherently spatial, and thus lends itself well to study through GIS. It is important that further development of improved tools for assessment of impacts of land use change on catchment hydrology takes place. It is a requirement that this development maintains a spatial framework.
This study has generated land capability and industry specific suitability maps extending across the ACT. This mapping has been used to calculate the potential streamflow reductions from increased intensive, high value agricultural industries in the study area. A significant reduction in streamflow has been estimated to occur should the maximum development scenario be realised. The results highlight the need for catchment or regional studies of this nature to indicate general trends resulting from management and policy decisions.
A number of limitations were discussed in the study, including the use of annual average streamflow totals for calculation of potential reductions and that only biophysical constrains were considered in generation of scenarios of land use change. Improvements to the methodology have been discussed, these include improved temporal modelling, increased economic and social analysis, closer interaction with industry groups and the use of flow indices to compare the impact of land use change scenarios.
The project has successfully integrated data derived from GIS analysis with hydrologic modelling and demonstrates some of the potential of these tools.
Croke, B.F.W. and Newham, L.T.H. (in prep). A System for Integrated Catchment Modelling (ICMS): Model Libraries. Environmental Modelling and Software, ICMS Special Edition.
Emery, K. (1988). Rural Land Capability Mapping, Soil Conservation Service of New South Wales.
Environment ACT (1999). Draft Water Resources Management Plan. Canberra.
Evans, J.P. and Jakeman, A.J. (1997). Development of a Simple, Catchment-Scale, Rainfall-Evapotranspiration-Runoff Model. MODSIM 97, International Congress on Modelling and Simulation, Hobart, The Modelling and Simulation Society of Australia Inc.
Gilmour, J.K. and Watson, W. (2001). Hydrological Economic Approach for Assessing Water Allocation Rules. In Proceedings of the Australian Agricultural Resource and Economics Society, 14th-15th January, Adelaide (CDROM).
Gilmour, J.K. and Croke, B.F.W (2000). An Integrated Modelling Approach for Assessing Water Allocation Rules. Hydro 2000, 3rd International Hydrology and Water Resources Symposium of The Institution of Engineers, Perth.
Jakeman, A.J., Littlewood, I.G. and Whitehead, P.G. (1990). Computation of the Instantaneous Unit Hydrograph and Identifiable Component Flows with Application to Two Small Upland Catchments. Journal of Hydrology, 117: 275-300.
Moore, I.D., Norton, T.W and Williams, J.E. (1993). Modelling Environmental Heterogeneity in Forested Landscapes. Journal of Hydrology, 150: 717-747.
Nix, H.A. (1986). A Biogeographic Analysis of Australian Elapid Snakes. Australian Flora and Fauna Series No. 7. Ed: Longmore, R., AGPS, Canberra.
Schreider, S.Y. and Jakeman, A.J. (1999). Land Use Changes and the Quantification of Farm Dam Development. In Impacts and Implications of Farm Dams on Catchment Yields. Ed: Schreider, S.Y. Canberra.
Sleeman, J.R. and Walker, P.H. (1979). The Soils of the Canberra District: Soils and Land Use Series, No. 58, CSIRO Division of Soils.
Strusz, L. (1979). 1:250 000 Geological Series – Explanatory Notes: Canberra. Australian Government Publishing Service, Canberra.
Vertessy, R.A. and Bessard, Y. (1999). Anticipating the Negative Hydrologic Effects of Plantation Expansion: Results From a GIS-Based Analysis on the Murrumbidgee Basin. In Forest Management for the Protection of Water Quality and Quantity, Second Erosion in Forests Workshop, Cooperative Research Centre for Catchment Hydrology.