Agricultural Research Centre, PMB 19, Mitchell Hwy, Trangie NSW 2823
Phone 02 6880 8000 Fax 02 6888 7201
Computer based pasture growth models are increasingly being used by land managers as a decision making tool. The current climate and pasture information combined with the predictive outlook for months ahead, give land managers a head start in planning future activities. Before these models can be put to use, a specialised process of calibration and validation must take place to make certain the model is producing accurate information. An essential component of the validation process is the collection of large amounts of spatially referenced data in the field. This data is compared with that produced by the model for the same period of time. Because the rangeland environment is so vast, traditional methods of collecting this type of data are both time consuming and expensive. With the use of a Global Positioning System (GPS), notebook computer and Geospatial software, multiple biological observations can now be made from a moving vehicle. Information on variables such as pasture biomass, tree/shrub cover, chenopod biomass, land condition and fire scars have been collected using this technique. The Aussie GRASS system is producing maps that reflect pasture condition in the rangelands. Land managers and agency staff in western NSW are currently using selections of these maps as part of a Monthly Seasonal Report that is under evaluation. These maps can help land managers, agribusiness and agency staff make timely, profitable and sustainable management decisions for the future.
Rainfall and pasture growth in Australia’s rangelands are highly variable in both space and time. Pastoralists must adjust their management to continuous changes in seasonal conditions. Similarly, agency personnel need to monitor land condition to properly administer the resource. Information relating to current climatic and pastoral conditions is a valuable instrument for decision making. Collection of the required data by traditional methods is both time consuming and expensive. Spatial modelling of pasture growth and related variables offers the potential to incorporate current information on regional seasonal conditions into decision making in a more cost effective manner. However calibration and validation of such models present significant challenges.
An essential component of the calibration/validation process is the collection of large amounts of spatially referenced data in the field. This data is compared with that produced by the model for the same period of time. To assist field data collection, a technique called spider mapping (Hassett et al. 2000) has been developed by researchers calibrating/validating the GRASP pasture production model for Queensland. The term spider mapping was used because the observations form an appearance very similar to that of a spider’s web when viewed on a map (see figure 1). Collaborators in the national Aussie GRASS, Australian Grasslands and Rangelands Assessment by Spatial Simulation (Hall et al. 2001) project have further developed the technique to calibrate/validate GRASP and several other models for their respective states. The technique uses a Global Positioning System (GPS), notebook computer and geo-spatial software to collect spatially referenced data from a moving vehicle. In western NSW the data collected to date have included estimates of pasture biomass, tree/shrub cover, chenopod sub-shrub biomass, land condition, fire scars and range type. These estimates or observations are input directly into the notebook computer by an experienced observer while the driver contributes by making estimates from their side of the vehicle. Up to 2500 observations over distances of 500km or more can be made daily using this technique. All pasture biomass estimates are calibrated with harvest cuts at regular intervals. Pasture cuts are taken over the range of biomass levels traversed to accurately represent each day’s observations.
To calibrate chenopod sub-shrub biomass, several pre-determined standard shrub sizes are used. When spider mapping in chenopod communities, observations include the standard shrub used, a correction factor for average shrub size and the sub-shrub density. The latter is calibrated using belt transects. Tree/shrub cover is recorded as % canopy cover of trees plus % canopy cover of shrubs. Estimations of % canopy cover along the same transects are made from an aircraft using a 200 x 200m area of assessment. The technique used from the moving vehicle is easily applied from a low flying aircraft. From the aircraft, sites showing relatively uniform tree/shrub cover over an area of approximately one square kilometre are selected and denoted as waypoints. On the ground, line transects are then used at these sites to measure % canopy cover, foliage projective cover and tree/shrub basal area in the shrub/tree community.
Figure 1 Foliage Projective Cover observations for spider mapping November, 1998.
In the pastoral areas of western NSW, collaborators in the Aussie GRASS project used a notebook computer linked to a Garmin 75 GPS for collection of spatially referenced data. The CIGS ©1998 (Geonautics International, Brisbane) GPS data acquisition application was used to facilitate the input of observations together with spatial coordinates. The software allows the operator to view geo-referenced topographic or satellite maps while inputting data. It is also able to overlay drainage systems, roads and cadastral boundaries onto the image. The current position of the vehicle is displayed as a red dot on the map and each observation logged into the computer is also displayed by a dot of user-defined colour and size (see figure 2). Function keys are configured according to the data needing to be recorded for that variable. All the observations are stored in an ASCII log file. The format of this file includes the date, time and spatial coordinate for each observation. The validation process takes the average value for a selected variable for each 5km by 5km grid cell estimated on the ground and compares with biomass values generated by the pasture growth model for the same 5km by 5km grid cell and the same time period.
Figure 2. CIGS © 1998 screen display showing logged observations made during spider mapping in far north-western NSW.
The Aussie GRASS project is an example of how a computer model can be used to produce maps that reflect pasture condition in the rangelands. Land managers and agency staff in western NSW are currently using selections of these maps as part of a Monthly Seasonal Report that is under evaluation. The report is based around 17 maps with a short description of each. Maps include the past months pasture growth in kg/Ha, pasture growth for the past month relative to historical records and total standing dry-matter (TSDM) for the past month relative to historical records. A map indicating the probability of exceeding median pasture growth in the following 3 months (based on analog years of the phase of the Southern Oscillation Index over the past two months), is included as a forecasting tool for users of the report. The maps are produced in full colour and information on Sea Surface Temperature (SST) and the Southern Oscillation Index (SOI) are included to give a fuller picture of current climatic events. These maps can help land managers, agribusiness and agency staff make timely, profitable and sustainable management decisions for the future.
Clipperton, S.P., Bean, J.M. (2000). A technique for rapid acquisition of spatial ecological data. In Proceedings of the Australian Rangeland Society Centenary Symposium, 21-24th August, 2000, Broken Hill, NSW, Australia pp. 175-177. Australian Rangeland Society.
Hassett, R.C., Wood, H.L., Carter, J.O. and Danaher, T.J. (2000). A field method for statewide ground-truthing of a spatial pasture growth model. Australian Journal of Experimental Agriculture, 40, 1069-1079.
Hall, W.B., Bruget, D., Carter, J., McKeon, G., Peacock, A. and Brook, K. (2001). Australian Grassland and Rangeland Assessment by Spatial Simulation (Aussie Grass). Final Report for the Climate Variability in Agriculture Program, April 2001.