Sub-Project 3: Information Systems
|
Regional Water and Soil Assessment for Mapping Sustainable Agriculture |
|
Sub-Project 3: Information Systems
|
|
|
|
The Information Systems Sub-Project will collate the results and present maps and modelling outcomes developed in the other two Sub-Projects on a consistent geographic base and undertake scaling from site to region. It will, in parallel with the Technology Transfer Sub-Project, establish communications with the users of the research outcomes. This Sub-Project will also integrate and present the regional results in space and time dimensions.
GIS and data integration systems for the LP, NCP, LPL, MLR and DT Catchment study areas, which build on the extensive work already done in both countries will be extended to support the new research as well as the management outcomes. This will integrate field, map, meteorological, hydrogeological and remotely sensed data, have a modelling capability and provide information for the decision support tools. Steps needed to complete the activity are:
First, related research on assessment, monitoring and scaling of regional water balance components and land degradation indicators using GIS and remotely sensed data will be reviewed. Then the remotely sensed data (e.g. TM and AVHRR images) and GIS themes (hydrogeology, meteorology, climatic averages, land use) not already in place will be collected for the study areas. A significant task at this stage will be to assess integrity of the data and establish the Data Quality Protocol which allows the reliability of separate and integrated data sets to be measured. These are needed for effective base data integration and geometric navigation. In China, data from the Chinese Meteorological Network and the Chinese Hydrogeological Survey are already well integrated and managed and available for this purpose.
Next, a generic GIS framework will be established which can be used for each of the focus areas (including the 38 N parallel of the NCP and the existing multi-scale GIS base at Yangling in China for the LP) to assist with the mapping and integration of information accessed by on-ground and remote sensing techniques. This framework will be based on a common, generic data model for the areas and applications addressed in both China and Australia and adhere (where possible) to data standards such as the Australian Spatial Data Transfer Standard which is currently being drafted.
To apply the up-scaling methods described in Section 2.1, the spatial strata used for mapping water and soils indicators using the remotely sensed data and target strategic sites will be identified in order to carry out the spatial processing needed for the Water Balance and Soil Environment Impact Sub-Projects. Crop stress and potential yield will be mapped at different scales and some of the site soils data aggregated to the same scales by spatial accounting. Then validated models from the Water Balance and Soil Environment Impacts Sub-Projects which extend the data from the focus catchment experimental sites to regional scales using remotely sensed data combined with GIS processing of data will be applied.
Finally, the GIS modelling techniques (e.g. using digital elevation models and derived indicators) and remote sensing information will be applied to the soil and hydrological data to extend the local soil and water models to predict water use and soil degradation at the regional scale through the base of standardised indicators. A validation phase will then be undertaken. It is important also for this activity to produce graphics that users can understand quickly and unambiguously.
Develop a generic water, soils, meteorological and hydrogeological GIS framework for the sites in both countries and use it as a base for data integration, upscaling and extrapolating to regions.
Establish appropriate methodologies for up-scaling, mapping and monitoring the water and soil indicators at regional scales.
Establish means to predict catchment-specific risks to sustainable and efficient water use due to land degradation through effective monitoring of the scalable indicators.
Integrate outputs from water balance modelling with remotely sensed data to monitor components of the regional water balance as well as potential yield.
Year 1 (97/98) Progress - as extracted from the Annual Report
There have been several review documents written in the first year of the ACIAR project. McVicar et al. (1997) discussed four recent applications that form the basis of knowledge and expertise introduced by CSIRO Land and Water for the ACIAR project. The four examples, briefly reviewed, were:
Techniques for using remote sensing to estimate crop yield have been reviewed by McVicar and Jupp (1998). This review focussed the current and potential uses of reflective, thermal and microwave remote sensing for assessing drought conditions. Low crop yield is one measure of drought and previous research was analysed in this comprehensive review.
Spatial data listings have been established to fulfill this milestone.
To make best use of daily meteorological data to provide necessary ancillary meteorological data to the surface energy balance models at the times of remotely sensed data acquisition which allow the models to be run. The minimum daily meteorological data set consists of daily maximum and minimum air temperatures and daily rainfall. From this minimum data set methods to derive air temperature, relative humidity and solar radiation at the AVHRR overpass times have been tested. The influence of wind speed has also been assessed. Finally, the sensitivities of these two methods are tested against intensive field data collected at the Yucheng Research Station, North China Plain.
The data sets that are required have been identified. These illustrate the design and population of GIS data layers as required for the ACIAR project.
As part of the project design ERDAS Imagine has been purchased for each site in Australia. Sites in China have access to this software as the selected catchments are CAS focus catchments.
The Adelaide Research group have made progress in establishing appropriate methodologies for up-scaling, mapping and monitoring soil indicators at regional scales. They have used satellite imagery to map land cover, topographic analysis to predict potential soil wetness and GIS analysis to combine information on roads, streams, and soil characteristics with the topographic and remotely sensed data. The aim of this work is to establish a means to predict catchment-specific risks to sustainable and efficient water use due to land degradation through effective monitoring of the scalable indicators.
The ISWC Information Systems Group have made considerable progress applying the indicator methodology to existing data sets and have generated several papers on the topic.
The SIAM Information Systems Group have been undertaking necessary staff development by doing GIS training courses and have been expanding their GIS and remote sensing data bases.
Year 2 (98/99) Progress - as extracted from the Annual Report
Strategic Scientific Research
One of the major stumbling blocks preventing thermal remote sensing from being operationally linked with regional energy / water balance modeling is the need for meteorological data at the specific time-of-day that the thermal remotely sensed data is acquired. There are many more stations recording daily meteorological data than stations recording data continuously e.g. well instrumented research catchments (which are costly and are usually for small catchments for short periods). Methods have been developed and tested which explore options for providing the required one-time-of-day meteorological data from standard daily data, this is in press in Agricultural and Forest Meteorology, see ref 29 (CBR_02). Overcoming this barrier allows thermal remote sensing to be operationally used with energy / water balance models over large areas like the Murray-Darling Basin and the North China Plain.
Designing and populating the GIS framework
Extensive time series of remotely sensed data has been obtained by CLW Canberra for all of China. 8 km2 PAL data from NASA for all of China set is from July 1981 until September 1994. The 1 km2 AVHRR data set from USGS is from 1 April 1992 until 30 September 1993 then from 1 February 1995 until January 1996, there is also some data for May 1996. These data sets have been obtained from the WWW and data processing including rectification to a common projection undertaken. Focussing on the NCP the integral under the NDVI time series curve is calculated for wheat growing period (day-of-year 59 to 171) and for corn growing period (day-of-year 181 to 293). Considering the temporal extent of the two data sets allows up-scaling to be performed focussing on the corn growing period in 1992, both wheat in 1993. If we slightly shorten the NDVI integration period by one month there is the opportunity to assess corn in 1993 and wheat in 1994.
Similar analysis will also be undertaken for the Loess Plateau, however there is extreme topographic control on both the at-satellite observed radiance and surface water flow. We lack a suitable digital elevation model (DEM) for the entire Loess Plateau to allow these effects to be normalised. Hence for the Loess Plateau this analysis is currently seen as exploratory. To assist in overcoming the lack of a suitable DEM for the entire Loess Plateau software to generate an accurate DEM, specifically ANUDEM has been purchased for ISWC. However, due to the time required for data capture considering the relative complexity of the Loess Plateau landscape it is unlikely that such will exist before this current ACIAR project is complete. The group in ISWC are concentrating on developing the indicators approach where they have suitable GIS data.
Up-scaling Soil Processes
The Adelaide Research group have made substantial progress in establishing appropriate methodologies for up-scaling, mapping and monitoring soil indicators at regional scales (Ref: 28 ADL_08). They have assessed and developed methodologies using soil processes data within a GIS framework to produce regional scale (ie. ~80 km2) assessments of drainage/waterlogging, salinity and soil acidity/alkalinity. These regional scale assessments are being verified by using detailed mapped information at catchment (2.0 km2 ) and toposequence (400 m within a 0.2 km2 key area) scales. Further funds have been made available from the National Land and Water Resources Audit (NLWRA) to augment this work. A number of techniques are being used to assess and predict poor drainage/waterlogging, salinity and soil acidity/alkalinity and produce a set of nested maps at a range of scales.
Our proposed methodology uses field pedology, vegetation, hydrology, topography and remotely sensed data within a GIS framework. This up-scaling methodology requires adequate data sets at multiple scales (see refs: 28: ADL_08; 38: CBR_03). A process-based approach has been used to characterise and assess natural resource status and condition. Recognition of soil process patterns at large/ point scale followed by controlled extrapolation of these patterns to smaller scales (catchments and regions) is essentially the approach adopted in this project. This technique relies on control data sets at the various scales to act as checks and or refinement tools in the up-scaling process.
This regional assessment has the potential to be used in conjunction with the published soil assessment manual to help landholders better map problem sites by recognising key field indicators and using the information as part of their property management planning. This approach will also increase landholder and regional adviser awareness of the extent of soil degradation through the use of field indicators (see refs: 35: ADL_09). The final report will be published as a CD-ROM/ WWW product and will provide significant outcomes from this project, which involved close collaboration between staff in CSIRO Land and Water, University of South Australia and Primary Industries and Resources, South Australia.
The ISWC Information Systems Group have made considerable progress developing data bases to rapidly survey soil erosion at a regional scale, see ref 25. (ISWC_04) and 26. (ISWC_06). Also several other papers that were discussed in the 97/98 annual report are attached, refer to 41. (ISWC_01), 42. (ISWC_02), and 43. (ISWC_03).
The SIAM Information Systems Group have been have been expanding their GIS, focusing on developing regional yield and meteorological data bases.
Year 3 (99/00) Progress - as extracted from the Annual Report
In CLW Canberra fundamental spatial research using a calculate-then-interpolate (CI) approach [in preference to interpolate-then-calculate (IC)] for regional hydrological modeling of moisture availability has been assessed. This research used AVHRR data in a novel way; as covariates for spatial interpolation. This technique inherently uses the high spatial density of remotely sensed data. This manuscript describing this strategic research has been invited into a Special Issue of Remote Sensing of Environment, refer to CBR_04. A companion Technical Report fully documents all technical details, including the basis for selection of spline interpolation models (CBR_05). For a daily time step regional water balance model this new CI method is vast improvement on the currently used IC approach, which relies heavily on the accuracy of interpolation of input driving variables.
Spatial Information Systems scientists in SIAM have been busy establishing data sets for the regional assessment of water use efficiency for the North China Plain from 1984 until 1996. This has taken large amounts of time and resources to acquire data for a large region in China. Developing these regional data bases has placed pressures on these scientists, which ‘point’ based scientists do not have to deal with. These scientists have performed a mighty task in establishing a regional data base for China, these efforts should not be overlooked. This work has started during this reporting period during a visit from Chinese staff to Australia (see Section 3.3 Travel and Meetings below) and the monitoring of water use efficiency for the North China Plain will be reported in next years annual report.
The ISWC Information Systems Group have made continued their considerable progress developing methods and the required data bases to rapidly survey land use, soil erosion at regional scales, see references ISWC_04 to ISWC_06. The methods and data bases established by this group over many years will be the basis of their research involvement, using the Loess Plateau as case study, for the “West-China Action Program of CAS” which underpins the “West-China Development Strategy”. This project was introduced in the Executive Summary, specifically the Future Research Activities section above.
In CLW Adelaide there have been large advances in the applied assessment of natural resources at regional scales. A methodology to construct soil-landscape models for regional assessment of susceptibility to drainage/waterlogging, salinity and acidity/alkalinity in an 80 km2 area in the Mt. Lofty Ranges has been developed (ADL_18; ADL_19). This method reduces cost and removes biases often introduced local experience, knowledge and personal judgement of the surveyor(s).
The methodology uses the field recognition of soil process patterns at point scale followed by controlled extrapolation of these patterns to catchment and sub-regional (80 km2) scales. Point scale data resulting from detailed investigations into pedology, hydrology, topography, geology and vegetation along a number of toposequence transects (approximately 400 m) within a 0.2 km2 catchment were compiled to provide information on representative landscape drainage/waterlogging, salinity and acidity/alkalinity patterns. This information was extrapolated to the catchment as a whole (2 km2) using information from a 1:5,000 scale soil survey. This involved the linking of soil and hydrological processes to the mapped soil units and the allocation of potential drainage/waterlogging, salinity and acidity/alkalinity classes to each unit. The next step was to extrapolate this data to the sub-region by integrating 1:50,000 scale soils information (obtained from PIRSA), Landsat TM, AIRSAR (ADL_16) and terrain analysis methods (e.g. ADL_20). These data were analysed within a raster GIS framework, using a linear weighted index modelling technique, reflecting the relative contributions of each data set. Representative areas of each mapped class of soil drainage/waterlogging, salinity and acidity/alkalinity in the catchment were then used as a training set to optimise the classification. The resulting data set was checked and verified by field survey techniques.
A number of nested map products including "best estimate" maps of drainage/waterlogging, salinity and acidity/alkalinity for 2 km2 (catchment) (sub-regional) areas (ADL_18) have been produced. Initial assessment of the results obtained show good qualitative agreement based on ground truthing from limited random site checks across the region (ADL_21). A set of data models, which defined the spatial analysis processes used in the project have been developed. Project staff acquired and installed the ANZLIC metadata entry tool software and compiled the required metadata set information in the final report (ADL_18).
The resulting data sets (ADL_18) give an assessment of degrees of potential soil drainage/waterlogging and salinity in the landscape for the 80 km2 sub-region. Approximately 11% of the area is classified as having the potential to be strongly waterlogged, with a further 27% potentially periodically waterlogged. Particularly noticeable are the indications of waterlogging moving up-slope in small gullies perpendicular to the principal drainage system. In certain locations these gullies link to patches of wet soils correlated with areas of groundwater discharge. The overall coincidence of areas mapped as potentially waterlogged in the landscape shows good qualitative agreement with ground truth data acquired from limited, random sites across the study area.
These maps were then used in conjunction with other landscape attributes (e.g. riparian vegetation, landcover and slope, etc) to produce a composite regional scale assessment of catchment condition (a rating from good to poor). A significant outcome of this work was the formalisation of metadata for current Mt Torrens data sets using the ANZLIC metadata entry tool software, which involved close collaboration between staff in CSIRO Land & Water, University of South Australia and Primary Industries & Resources, South Australia. Results of this work were presented at several National and International conferences and field trips and have been published (ADL_18).
Extension of the catchment condition for work has taken place in the upper Torrens region where 230 catchments have now been rated by the same techniques previously applied to the 55 catchments in the Mount Torrens (80km2) region (ADL_17).
The office assessment using GIS (ADL_18; ADL_17) should first be used to broadly identify the regional extent of land degradation as a precursor before applying the practical on farm “Soil-landscape and vegetation field key” developed by Fitzpatrick, Cox, and Bourne in 1997 (Catchment Management Series; CSIRO Publishing). The field key will help landholders to map problem sites (“hot spots”) on aerial photographs and assign practical solutions. Results of this work were presented at the 17th Biennial conference of the Australian Clay Minerals Society Inc. 9-14th April, 2000 (ADL_19).
As part of the investigations into soil moisture estimation using radar remote sensing, improved algorithms were presented at the “Pacific Rim AIRSAR Significant Results and Planning Workshop” from 24-26 August, 1999 (ADL_16). The waterlogging assessment work was submitted in March, 2000 for inclusion in the 10th Australasian Remote Sensing and Photogrammetry Conference. We have tested the utility of developed spatial models for waterlogging and pH properties using field site data for the above attributes (ADL_18, ADL_21). This work is to continue, potentially linking with a NLWRA proposal to translate the spatial methodologies to Victoria and Queensland during 2000 and 2001.
Year 4 (00/01) Progress - as extracted from the Annual Report
Please direct research and general email enquiries to
Tim McVicar in Australia and Yang Qinke (ISWC) or Zhang Guanglu (SIAM) in China
Last updated 7 May 1999 communicator@eoc.csiro.au
Copyright 1999, CSIRO Australia
Use of this web site and information available from it
is subject to one or more of the following leagal notices
CSIRO Legal Notice and Disclaimer