|
Spatial Query Tool to estimate water use for key land use types
3.0 Methodology The project methodology shown in Figure 1 is based on a water fall model; the output of one process is an input for the other process. The methodology is subdivided into four major processes as spatial data input and analysis, SoilFlux model, grid model and query model, which are explained in the following sections. Figure 1 Methodology 3.1 Spatial data input and analysis The study focuses on areas of private land within the state because many areas of public land are reserve areas which are protected from land use changes. Where plantation forestry activities occur within public land areas, it was assumed that future land use change would be minimal. The study only considered dryland (i.e. non-irrigated) land uses. Based on land use mapping data, an understanding of the land uses likely to have a hydrologic impact, and land uses that may experience growth or reductions in the future, the following land uses were selected as the focus for this study:
The state was divided into 1 km2 grid cells and a value for each input (see table 1.0) was determined each cell. All the unique combinations of rainfall interval, soil, geology, and depth to water table across the state were determined, and only these combinations were modelled. 3.2 SoilFlux Model The modelling of each land use required approximately 11,000 model runs. The model was run on a daily time step over the period of 1950 to 2007. The annual results over the period, 1960 to 2007 were averaged to provide a single mean annual value of:
 Table 1.0 Spatial data sets used in the study The SoilFlux model developed at SKM, considers the vertical movement of water and solute in the unsaturated zone of soil and surface geology and to and from a saturated zone (Daamen et al., 2001a). The model was developed in Fortran software. SoilFlux uses Richard’s equation for simulating water movement. The plant water use was modelled using SoilFlux which is a 1-dimensional model that describes water movement through the soil profile. Lateral flow to groundwater systems or streams is not modelled although it is estimated from water flow out of the model. SoilFlux was selected as the modelling tool for this study for the following reasons:
3.3 Grid Model The output of the SoilFlux modelling was linked to the 1km2 spatial grid as shown in figure 1. The basic idea behind the grid model is that it is an easy to manage from a database perspective. The grid acts as an input to the development of the spatial query tool.   Table 1.0 Spatial data sets used in the study 3.4.1 Introduction The land use change maps and the outcomes of the SoilFlux modelling have been incorporated into a spatial query tool (see Figure 2). This tool allows the user to interrogate the results of the modelling, and report on the likely impacts of a particular land use change at a catchment scale. It also allows the estimated impacts of land use change to be placed in the context of water resource availability in a particular area (see Figure 3). The spatial query tool provides the user with a report for each query (see figure 7, 8), which details the likely impacts in terms of plant water use and change to surface water availability. The report also contains information on the Sustainable Diversion Limit for the catchment, as well as critical downstream catchments, in order to fully understand the impacts of the land use changes. The query tool also has the functionality to query multiple land use change (maximum of three at a time) at a catchment level. The following sections discuss the types of queries, design of query tool and the query results.  Figure 2 Spatial Query Tool 3.4.2 Types of queries Sample query type 1 (Landuse change query): The Figure 3 is the query interface for sample query type 1. For a selected catchment, if the landuse is changed from Annual Crop to Native grassland, then, what is the Change in annual Evapotranspiration (ET)? what is the Change in winterfill ET ? what s the Change in non winterfill ET? what is the Change in annual surface water volume? what is the Change in winterfill surface water volume? what is the Change in non winterfill surface water volume? what is the Change in annual deep aquifer drainage? what is the winterfill surface water volume and consumptive entitlement allocated? does the SDL exceed by landuse change? Yes /No what is the new volume available?  Figure 3 Impact of land use change query (sample query type 1) Sample query type 2 (Query by rainfall zone) : The Figure 4 is the query interface for sample query type 2. For a selected rainfall district, min rainfall, depth to water table, geology, A Horizon, B Horizon, Ak_sat, B_Ksat, and if the landuse is changed from annual crop to native forest, then what is the Change in annual Evapotranspiration (ET) ? what is the Change in winterfill ET ? what is the Change in non winterfill ET?  Figure 4 Query by rainfall zone (sample query type 2) 3.4.3 Design of Spatial Query Tool The major strength of the spatial query tool is its functionality to perform scenario based queries. Standard GIS software has limitation in performing scenario based queries; hence customising the GIS software to handle scenario based queries was essential. A key challenge in designing a customised tool bar for querying in GIS is the large volume of spatial data. The usability of the tool is affected if the user must wait for a prolonged period of time for the display of the results. According to industry standards ~1 minute time is the standard wait period for a user to display the results of interactive query. Therefore effective query optimisation techniques were adopted in the design of the spatial query tool. The spatial table CELLS consisting of 220,000 polygons as a grid was the outcome of the SoilFlux modelling described in section 3.3 is the layer used by the spatial query tool for scenario based queries. Figure 5 shows the query optimisation adopted in the design of the query tool. The non-optimised query performs the calculations on all the 220,000 cells and then aggregates the results for the selected catchment. As a result the process takes about 20 minutes. The optimised query first selects the catchment cells, based on the selection set, the calculations are done on the fly and the results are aggregated for output. The total time taken for execution is only 5 seconds. Note a standard Windows PC with Intel Core 2 Duo processor with 2 GB Ram was used for the implementation.  Figure 5 Query optimisation for query1 For query type 2 (query by rainfall zone), a drill down query approach was adopted as shown in figure 6. In a drill down approach, the user is only allowed to query from the data available from the database. In the drill down approach, the user first selects the rainfall zone (see figure 4). Based on the rainfall zone selected, the minimum rain data is populated in the drop down for the user to select. Likewise the other parameters are listed based on the previous selection. The advantage of the drill down approach is that the user is prevented from making any illogical queries. The query execution time is faster as there are only a few records to be processed as a result of filtering the records in each drop down selection.  Figure 6 Drill down query approach adopted for query type 2 3.4.4 Query results The results of both queries query type 1 and query type 2 were designed to be reported as spatial and non-spatial data. Figure 7 shows the results of sample query type 1. The spatial window in the report shown in Figure 7 highlights the selected catchment and it’s critical downstream. The input parameters of the query are reported in the first section of the report as: Basin number, critical downstream, etc. The next section in the report, SDL CATCHMENT WHERE LANDUSE CHANGE IS OCCURING, is reported from the CELLS (grid) table directly. The section, POST LANDUSE CHANGE INFORMATION in the report, is the information generated on the fly based on calculations. The last section of the report consists of information on the critical downstream. Furthermore, the spatial query tool has the functionality to export the layout as a pdf. The advantage of exporting the query results to the pdf is:
 Fiqure 7 Results of query type 1, processed in 5 seconds  Figure 8 Results of query by rainfall zone (sample query type 2, processing time 4 seconds) 4.0 Summary and Conclusions Given the simplifications in the modelling approach, there is considerable uncertainty in the results at any specific location but we expect estimates of changes to be reasonable when considered at a catchment scale and comparable between catchments across the state of Victoria. The results are useful for obtaining preliminary estimates of changes in ET and surface water availability resulting from changes in land use. They can also be used for identifying priority areas for more detailed modelling based upon consideration of the current hydrologic stress and anticipated changes in land use. The following recommendations and conclusions are summarised as follows:
The authors express their sincere gratitude to Department of Sustainability and Environment, the Victorian government department for funding the project. The authors also thank their colleagues Georgina Race, Peter Hill, Rachel Murphy and Carl Daamen for their valuable inputs and review of this paper. 6.0 References
Page 2 of 2 | Previous | |