Previous PageTable Of Contents

Improved management of the Jemalong Irrigation District by remote sensing and GIS

Ian McGowen1, Sally Duff2 and Ian Smith3

1NSW Agriculture
161 Kite Street (Locked Bag 21), Orange NSW 2800
2
Jemalong Irrigation Ltd,
PO Box 520, Forbes NSW 2871
3
NSW Agriculture,
Camp Street (PO Box 369), Forbes NSW 2871
1
Corresponding author – Phone: 02-63913195, Fax: 02-63913767
Email: ian.mcgowen@agric.nsw.gov.au

Abstract

A regional Geographic Information System (GIS) has been used as a means to collate, store, integrate and analyse data about the environment, farming practices and infrastructure over the Jemalong Irrigation District. The GIS has allowed the integration and analysis of a wide range of data derived from ground based, airborne and satellite remote sensing systems, as well as cultural and land use data information obtained from farm surveys and Landcare groups. Information has also been collated from Government agencies and reports by consultants.

This information has been used for the production of a Land and Water Management Plan for the district. The GIS has proven integral to the development of the plan, through provision and analysis of information on soil type, land use, irrigation layout and water table levels. These data have been used to identify developing problems, and to target and plan amelioration through management changes.

The GIS will continue to be used as a planning and monitoring tool for the implementation of the Plan. For example, all water ordered by landholders in the District is now based on a GIS paddock identification number. In the future, information will be collated on the amount of water applied to each paddock and the yields achieved. This will allow estimation of water use efficiencies for various crops over a range of soil types, and potential accessions to the water table. It will be used for monitoring Plan targets, identifying priority areas for rehabilitation, to allow the sustainable management and development of the District.

Introduction

The Jemalong Irrigation District, covering 93,000 ha, is located in Central Western NSW between Forbes and Condobolin. Its boundaries are the Lachlan River to the north, Lake Cowal to the south and narrow hilly ranges to the east and west.

The district is primarily a mixed farming area, with a rainfall of 432mm/annum and enterprises including prime lamb and cattle production, irrigated and dryland summer and winter cropping, irrigated lucerne and lucerne seed production. Between 12,000 – 20,000 hectares are irrigated annually (Anon, 1994a) from a current total irrigable area of approximately 41,500 ha. Over the 1998/99 irrigation season, 16,400 ha were irrigated.

Rising groundwater levels over the district present a major threat to both dryland and irrigated agriculture. Approximately 32,000 ha of the district are predicted to have a water table at 2 m or less by 2025 (van der Lely, 1997). To address this and other issues relating to the sustainable management of the area, funding was sought for the creation of a Land and Water Management Plan.

As part of this process, a GIS was set up for the area. This allowed data essential for development of the Plan (and for farm management) to be captured and analysed. The GIS allowed the implications of various land management options to be modelled, and provided a source of information for farm and catchment level planning by local Landcare groups. The GIS is now being used to assist in the implementation of the Land and Water Management Plan targets, and for Landcare Groups and individual landholders to plan and implement management strategies.

Data layers and their source

A diverse range of data were obtained for use in the GIS, and new data are continually added.

Included are data captured and collated from:

  • ground based, satellite and airborne remote sensing systems;
  • Government records and plans;
  • reports and surveys by Government Departments and contractors; and
  • cultural, land use and other data provided by landholders and Landcare groups.

Property and paddock boundaries, land use and irrigation

The most important data set held over the Jemalong Irrigation District is land use, which includes detailed information for every paddock in the area since 1992, and for areas of the district since 1990.

The land use GIS layer was created from digital cadastral (parish and portion) data, provided by Land and Property Information NSW. From these data, property boundaries were selected, by comparison with hard copy farm plans. These boundaries were overlayed on a high resolution (10m pixel) SPOT panchromatic satellite image in the GIS. Using the farm plans, paddock boundaries were then identified on the image and digitised, to form a paddock boundary layer. Data on ownership, irrigation entitlement, crop or pasture type, irrigation layout, irrigation type and reuse systems for each paddock were then entered into the GIS (McGowen et al., 1996). This information and is updated yearly through landholder survey.

Remnant woody vegetation

Information on the extent and location of remnant woody vegetation over the district was required for modelling of water use across the catchment. This was important for the ‘do nothing’ scenario in the Land and Water Management Plan, to predict likely water use, increases in water table levels and potential land degradation.

The remnant vegetation study demonstrated the advantages of a GIS approach, as initial estimates of remnant woody vegetation cover across the district were only 3% ground cover and provided no information on the location or density of the vegetation. To provide a more reliable estimate and to give information on the spatial distribution and density, Landsat Thematic Mapper (TM) imagery was classified into three levels of woody vegetation cover. This provided a better estimate of 10%, however the scale and resolution of the imagery did not allow accurate mapping of scattered woody vegetation, and the available imagery covered less than 80% of the district (McGowen et al. 1996). A final survey was carried out which combined analysis of aerial photographs with field checking. The resulting maps were incorporated into the GIS and gave an improved estimate of 17% and some identification of species, although no assessment was made of density (Cunningham, 1998).

The updated remnant woody vegetation cover information was integrated with district soils data in the GIS. The Landsat data provided information for economic modelling to estimate the agricultural impacts of soil salinisation over the district (McClintock and Jones, 1995), and the data derived from the final survey is being used for better directed planning of tree plantings.

Soil maps

Four soil maps are held over the district. These include a Department of Land and Water Conservation soil landscape map, although at 1:250,000 scale this map is too broad for all but general use (King, 1998). Two basic soil maps are currently in use, one with general soil classes (Kelly, 1971) and another compiled by advisors and landholders, based on soil texture and mapped at 1:50,000 scale (McGowen et al., 1996). The latter has had the most use, despite the limitations of uncertainty on boundaries and the description of the classes. A major difficulty with both soil mapping and management in the district is the extreme variability in soils, in both physical and chemical characteristics. Given this variability, an attempt to map soil classes more accurately was made through the analysis of airborne radiometric data, although this has proven of limited value (Anon, 1995). Despite the limitations of all the soil maps, they form an important base layer and are used in many GIS analyses.

Infrastructure

Irrigation infrastructure data includes the major irrigation supply channels, farm channels and supply outlets. For the supply outlets, data are recorded on the size (capacity/type), channel, property and paddocks serviced, and the yearly throughput of water.

Water table, salinity and vegetation vigour data

Historical and current data from a network of piezometer and water table wells across the district are held in the GIS, and are updated regularly. The piezometer and water table well data indicate the level of the water table in confined and unconfined aquifers respectively, and have been collected since the late 1960s. Other data include regular groundwater salinity readings, as well as soil colour, texture and pH data collected during well construction (McGrath, 1994, 1997).

Neutron moisture meter data along the Warroo main supply channel between 1992 and 1993 have also been collected and entered into the GIS, as a means of determining the influence of channel seepage, rainfall and irrigation on groundwater levels (Smith and Rose, 1993).

Broadscale electromagnetic (EM31) survey data, conducted over part of the district, were used to determine current and developing salinity problems and to predict future scenarios. Although these data were coarse, they were entered into the GIS and surfaced to identify problem areas (Williams and Williams, 1992; Williams, 1993; McGowen et al, 1996).

A sequence of 7 Landsat Thematic Mapper (TM) images spanning a 2.5 year period were used for assessment of vegetation vigour, salinity and waterlogging over the district. The normalised difference vegetation index (an index created from the near infrared and red bands of the image) was calculated for each image, and areas consistently low in vigour across the sequence of imagery were identified. The results were correlated with the EM31 survey data, and problem areas investigated to determine the reason for their low vigour (McGowen et al. 1994, 1996; McGrath, 1997).

Due to the development of salinised and high water table areas in the Jemalong Irrigation District, the National Landcare Program selected the area for the evaluation of airborne electromagnetic (EM) and magnetic survey. Data were collected in various ‘channels’ to provide information on likely salinity at a range of ‘depths’ in the soil profile, and were validated by intensive EM surveys, the data from the broadscale EM31 survey and the Landsat vegetation greenness data (Anderson and Street, 1993; Odins et al., unpub. data). The survey found that there was a large amount of variability in surface conductivity across the district, which was not adequately detailed in the broadscale EM31 survey. The airborne EM survey proved useful in understanding the distribution of conductivity/salinity within the soil profile, movement of saline groundwater and catchment level salinity monitoring.

These data, together with hydrogeological survey information, have been used to predict future water table levels given a scenario where no changes are made to irrigation practices (Williams, 1993, 1995; van der Lely, 1997).

Channel seepage zones

The Warroo main supply channel loses 3,000ML/year through seepage into permeable soils associated with prior stream networks along the channel (van der Lely, 1993). Identifying these zones is of critical importance to reducing accessions (losses) to the water table and preventing the onset of salinity and waterlogging problems in the area. Surveys were carried out to investigate seepage, using data from seepage meters within the channel, EM31 survey, water table well and neutron moisture meter data. These data were combined with information from landholders and advisors to form a seepage zone map (van der Lely, 1993; Smith and Rose, 1993). These predictions were complicated by the effects of rainfall, flooding, irrigation and changes in the amount/depth of water within the channel. The information suggests that while seepage does occur from the channel, the effects are generally seen some distance away. Further work to identify the actual seepage zones, and to determine areas to which seepage water is moving has involved the use of airborne thermal imagery, captured at night and combined with Landsat TM imagery (McGowen et al., 2001).

Other data

A large range of ancillary data have been added to the GIS. These include the location of floodways, drainage basins, wetlands, banks/levees, and windbreaks/tree lots.

How are these data being analysed and used?

Water use efficiency, allocation and planning

Irrigators purchase water by quoting their GIS paddock number, so information can be kept on water use by paddock. Eventually, information on the crop yields will be kept, so that estimations can be made on water use efficiency over various soil types. This will allow information on losses to the water table to be calculated. However, there are some difficulties in this assessment as a single supply outlet often supplies multiple paddocks. Whilst water may be ordered for a particular paddock, it is often split between a number of paddocks.

Once allocations are announced each year, irrigators are requested to supply an estimate of their expected water use per paddock, and the type of crop to be grown. This information is used to determine water demand and the capacity of the supply channel network to cope with the demand, given the likely irrigation timing for the crops and the area to be irrigated. The switch to high value summer crops such as maize and rice is putting increased pressure on the supply network, and the predictions are used to indicate which channels may need upgrading.

Water table and seepage predictions - influence on land use change

The water table predictions for the district are that by 2025 nearly 32,000 ha are at risk of developing a water table of 2 m or less. Integration of the prediction maps with the land use/irrigation data has identified that 430 paddocks laid out for irrigation (45% of the total, or approximately 19,400 ha) either wholly or partially fall within the predicted shallow water table zone (Figure 1). Of these, less than half are landformed or use spray or micro irrigation systems (209 paddocks, or approximately 7,850 ha). The remainder therefore represent the highest priority for landforming.

The need for more efficient use of water and sustainable land use, in part to prevent increases in the water table, has been recognised in the Land and Water Management Plan targets. The targets have had a great deal of support from local Landcare groups and landholders, who recognise the need for such changes. Some of the Plan targets, and their current level of achievement based on 1998/99 GIS data (as 1999/2000 data collection is not yet complete) are:

  • To have 90% of delivered water supplied to landformed surface irrigation layouts, rice areas, sprinkler or micro irrigation systems by 2010. Currently, 76% of delivered water is now supplied to such areas (which represent 47% of all irrigable paddocks).
  • To have 50% of delivered water supplied to paddocks that can be serviced by a recycling system by 2006, and 90% by 2008. Currently, 68% of delivered water is supplied to such paddocks (which represent 43% of all irrigable paddocks).
  • To have 50% of delivered water supplied to properties with a farm plan by 2005 and 90% by 2007. Currently, 46% of delivered water is supplied to properties with a farm plan.

Integrating the water table prediction map with the soils and land use data also allowed identification of areas of annual pasture growing in non-irrigated paddocks on light soils (sands, sandy loams and loams). These potentially contribute major accessions to the water table, and represent a high priority for change to high water use deep-rooted perennial species such as lucerne (Figure 2). Another Plan target is to replace 10,400 ha of annual pastures with perennial species by 2015. This includes replacing 1,400 ha of dryland annual pastures on light soils. Currently, there are 9,100 ha of dryland annual/natural pastures in paddocks which have (or contain some) light soils, and have not been used for cropping since 1992. These represent the highest priority for resowing. In comparison, there are 5,100 ha of perennial pastures and trees on light soil areas. Also, approximately 5,900 ha of other dryland annual pasture paddocks on light soil areas are regularly used in a rotation with crops.

Combining the land use data with the Warroo channel seepage zone has identified areas for priority sowing to deep rooted perennial plants. Sowing to such species will allow better use of seepage water, minimise accessions to the watertable and reduce the onset of salinity and waterlogging. The data indicate the good landholder response to this information, with an increase in the area sown to lucerne in these zones from 625 ha in 1990 to 846 ha in 1993 (McGrath, 1994), increasing to 1,445 ha in 1999. When other perennial pasture species were included, the total area sown to deep rooted perennials increased to 2,489 ha in 1999. This represents approximately 53% of the estimated seepage zone, but falls short of the local Landcare group’s plan of 75% of the seepage zone sown to such species by 2000/2001 (Anon, 1994b). However, the expected stand life for lucerne in the area is about 5 years, and at any one time a number of paddocks will be rotated out of lucerne and cropped for some years prior to resowing.

Figure 1: Irrigable paddocks with a high priority for landforming (in yellow), Jemalong Irrigation District

Figure 2: Annual and perennial pastures on light soils, Jemalong Irrigation District

Currently, airborne thermal and radar imagery are being analysed to attempt to better define the location of seepage areas adjacent to the supply and on-farm channel systems, and to identify priority areas for channel sealing. The analysis of these data suggests that major leakage from the Warroo main supply channel is occurring where the channel intersects and runs adjacent to or along the prior stream (old river) network and the associated light textured soils. Deep rather than shallow seepage appears to be occurring along large reaches of the channel, with water movement into and along the old river beds associated with the prior stream network (Figure 3) (McGowen et al., 2001). Only one area of surface seepage was identified in a small zone adjacent to another major supply channel. This confirms the findings of Smith and Rose (1993), who found most evidence of seepage some distance from the channel rather than adjacent to it. Although seepage varies with rainfall, irrigation practices, time of year and the amount of water in the channels, the analysis of the thermal imagery is reasonably consistent with seepage locations identified by the limited previous surveys and local knowledge. However, the major disadvantage is that it provides information on where seepage is occurring, rather than information on the extent of water loss.

Figure 3: Seepage/prior stream locations mapped on thermal imagery

Salinity risk

Analysis of salinity risk has been carried out through combining the water table data, the water table risk maps, and the airborne and ground based EM surveys. The analysis identified that areas at risk from salinisation generally occurred adjacent to one particular soil type, the ‘red clay loams’ (from the landholder soil map). This information was provided to local Landcare groups to assist in planning to reduce the onset of salinity, and has been used in conjunction with other data on channel seepage (McGrath, 1994).

Tree corridor planting

The remnant vegetation maps are currently in use by Landcare groups to better target windbreak and tree lots. Attempts are in progress to join isolated patches of remnant vegetation with tree corridors, to allow movement of native fauna (Figure 4). Also, tree plantings are being targeted to areas of light soil between existing remnant vegetation stands, to reduce accessions to the water table in these areas. As part of the Land and Water Management Plan targets, 300 ha of new trees are to be planted and 2,800 ha of existing remnant vegetation fenced off by 2015. This will assist in preventing further loss of remnant vegetation and encouraging regrowth.

Figure 4: Directed windbreak/tree lot planting to join up areas of remnant vegetation.

Conclusions

The Jemalong Irrigation District GIS has been an integral part of the development of the Land and Water Management Plan, and will continue to play an essential role in its implementation. The use of the GIS has allowed more reliable analysis of data across the district, allowing existing and developing problems to be identified. Consequently, it has allowed better directed planning and management changes to ensure sustainable management of the irrigation district.

Acknowledgements

The assistance of the landholders, advisers, Landcare groups, Land and Water Mangement Plan officers, Jemalong Irrigation Ltd staff and the Department of Land and Water Conservation in providing information for the project is greatly appreciated.

Land and Property Information, NSW provided digital cadastral data for the project under a special licence agreement. The National Landcare Program provided funding for the airborne electromagnetic survey.

The airborne EM survey was conducted by World Geoscience Corporation Ltd, airborne thermal data was acquired by Air Target Services and airborne radar data was acquired by NASA.

Digitising, data entry and image georeferencing were conducted by Bev and Joanne Morcombe, Gillian McRobert, Catherine Cook and David McGowen.

References

Anderson, H. and Street, G. (1993). Interpretation report, Questem salinity survey, Jemalong-Wyldes Plains, NSW. World Geoscience Corporation Ltd, Floreat, WA.

Anon (1994a). Water Wise Farming. A discussion of Land and Water Management Issues for the Jemalong-Wyldes Plains Catchment. Jemalong-Wyldes Plains Community Steering Committee, Forbes, NSW.

Anon (1994b). Warroo Landcare Group Land and Water Management Plan. Warroo Channel and Water Table Management Association Inc.

Anon (1995). Soil descriptions for the Jemalong/Wyldes Plains using Airborne Radiometric Data - Final Report prepared for NSW Agriculture. Environmental Research and Information Consortium Pty Ltd, Canberra, ACT, Australia.

Cunningham, G. (1998). Remnant vegetation survey – Jemalong Land and Water Management Plan Area. Geoff Cunningham Natural Resource Consultants, Sydney.

Kelly, I. (1971). Jemalong and Wyldes Plains Irrigation Districts: Geomorphology, Soils and Groundwater Conditions. NSW Water Conservation and Irrigation Commission.

King, D.P. (1998). Soil Landscapes of the Forbes 1:250,000 sheet (Condobolin- West Wyalong – Parkes – Grenfell). Department of Land and Water Conservation, Sydney.

McClintock, A. and Jones, R. (1995). Agricultural impacts of soil salinisation and waterlogging in the Jemalong Irrigation District: an economic evaluation of the “current trends” scenario. A report to the Steering Committee of the Jemalong Land and Water Management Plan. Economic Services Unit, NSW Agriculture, Orange, NSW.

McGowen, I., Dudgeon, G., and Brandsema, F. (1994). Salinity detection using Landsat Thematic Mapper imagery to measure plant vigour. NSW Remote Sensing 5 (2), 28-29.

McGowen, I.J., Dudgeon, G.S., Smith, I., McGrath S. and Brandsema, F.A. (1996). Remote sensing and GIS for better land management in the Jemalong Irrigation District and surrounding areas. In Proceedings of the 8th Australasian Remote Sensing Conference, 22nd – 26th March 1996, Canberra, ACT, Australia (CDROM).

McGowen, I.J., Duff, S.L., and Smith, I. (2001). Identifying channel seepage using pre-dawn thermal imagery. In Proceedings of the IEEE 2001 International Geoscience and Remote Sensing Symposium, 9th – 13th July 2001, University of NSW, Sydney (CDROM).

McGrath, S. (1994). Resource Information from the GIS. A report on soils, cropping, irrigation and watertables for the Warroo Landcare Group Area, 1990 – 1993. NSW Agriculture, Forbes, NSW.

McGrath, S. (1997). The Geographical Information System. A report on the GIS used in the production and implementation of the Land and Water Management Plan. NSW Agricuture, Forbes, NSW.

Odins, J., O’Neill, D.J., Beckham, J., McGowen, I. and Nicholson, A. (1996). Validation of airborne EM survey over Jemalong/Wyldes Plains area of NSW. Draft Report (unpublished), Technical Services Division, Department of Land and Water Conservation, Parramatta, NSW, Australia.

Smith, I. and Rose, K. (1993). Jemalong and Wyldes Plains Irrigation District Land and Water Management Plan – An investigation into the effects of channel seepage, rainfall and irrigation practices along the Warroo main channel. NSW Agriculture, Forbes.

van der Lely, A. (1993). Channel seepage from the Warroo main canal, Jemalong Irrigation District. Technical Report No. 93/04. Department of Water Resources, NSW.

van der Lely, A. (1997). Jemalong Land and Water Management Plan – Soil salinity assessment based on the regional groundwater balance. Department of Land and Water Conservation, NSW.

Williams, B.G. and Williams, A.F. (1992), Electromagnetic induction survey of the Jemalong Plains Area.

Williams, B.G. (1993). The shallow groundwater hydrology of the Jemalong-Wyldes Plains irrigation districts.

Williams, B.G. (1995). The effect of a do-nothing option on future salinity hazards in the Jemalong-Wyldes Plains area.

Previous PageTop Of Page