(i) Purpose and context of the project

Desertification of grazing lands in the world's arid regions is an issue of major significance to scientists and policy makers and yet knowledge of its extent is often only available after degradation is either impossible or uneconomic to repair. Similarly land users often find it difficult to detect the effects of degradation before productivity loss is very significant.

In the more arid part of western Rajasthan, annual rainfall is predominantly monsoonal and highly variable. Grazing is the major land use although cropping and vegetation cutting are also practised by subsistence farmers. Desertification is occurring at an alarming rate as a consequence of overuse. Unfortunately, rainfall variability causes extreme variability in production year to year, so that the "noise" associated with seasonal conditions masks the decreasing trend in productivity.

This project aims to allow better discrimination between erratic rainfall and grazing, as the two causes of variability in production in Rajasthan, and thus to more effectively diagnose desertification. The use of remote sensing will allow this to be achieved economically in extensive areas of low productive potential.

(ii) Names of collaborating research institutions

CSIRO Centre for Arid Zone Research, Alice Springs, Australia (CSIRO)
Central Arid Zone Research Institute, Jodhpur, India (CAZRI)

(iii) Results or expected results and why these are believed to be important/valuable
ACIAR's approval to commence was given in September 2000, to begin formally from July 2000. Two CSIRO scientists and an ACIAR representative visited CAZRI in March 2001 for the first project team meeting. Progress with site selection and on-going literature searching was reviewed, essential facilities were assessed, a field visit and preliminary training was undertaken and staffing and logistics were discussed. Specifications and costings were developed for equipment and a GIS facility.
Following a visit to the Indian Council for Agricultural Research (ICAR), it became clear that the formal Letter of Sanction required by ICAR had not been obtained, and project expenditure could not proceed. Negotiations regarding intellectual property which had been initiated in February 2000 were revisited during April/May 2001, and the Letter of Sanction was signed by the Secretary of ICAR on 21st June 2001.
Letters of invitation to two CAZRI scientists to visit CSIRO for essential training in remote sensing and field technologies were developed in anticipation. Training and equipment were required urgently for fieldwork before the end of Rajasthan's monsoon in August and Australia ordered equipment on India's behalf so that it would be available in time.
Further progress before 30th June 2001 was not possible, but both equipment purchases and training were expected to proceed fairly smoothly subsequently. These were crucial to the development of a vegetation cover index in Rajasthan, for interpreting archival satellite imagery as well as imagery to be obtained in real time during the remainder of the project.

(iv) Likely direction of future research activities.

The project is now on track to achieve its intended outcomes. A training visit by two CAZRI scientists to Australia in August will ensure that fieldwork in Rajasthan can proceed as planned. This fieldwork will initially resolve how to gain entry to village communities and then include gathering of biophysical, livestock and socioeconomic data to assist the interpretation of remotely-sensed data. At the end of our first year (2000/01), our research directions remain as originally proposed. In addition to the outputs already mentioned, we aim to develop a robust index of vegetation cover derived from satellite data, a GIS for the area, images and indices to interpret the extent of human impact, products for explaining the technical outputs of remote sensing in ways easily understood by village people, and physical outputs including improved infrastructure, trained staff and standardised procedures for image processing.

(i) Purpose and context of the project

Desertification of grazing lands in the world's arid regions is an issue of major significance to scientists and policy makers and yet knowledge of its extent is often only available after degradation is either impossible or uneconomic to repair. Similarly land users often find it difficult to detect the effects of degradation before productivity loss is very significant.

In the more arid part of western Rajasthan, annual rainfall is predominantly monsoonal and highly variable. Grazing is the major land use although cropping and vegetation cutting are also practised by subsistence farmers. Desertification is occurring at an alarming rate as a consequence of overuse. Unfortunately, rainfall variability causes extreme variability in production year to year, so that the "noise" associated with seasonal conditions masks the decreasing trend in productivity.

This project aims to allow better discrimination between erratic rainfall and grazing, as the two causes of variability in production in Rajasthan, and thus to more effectively diagnose desertification. The use of remote sensing will allow this to be achieved economically in extensive areas of low productive potential.

(ii) Names of collaborating research institutions

CSIRO Centre for Arid Zone Research, Alice Springs, Australia (CSIRO)
Central Arid Zone Research Institute, Jodhpur, India (CAZRI)

(iii) Results or expected results and why these are believed to be important/valuable

The project has made advances in all its objectives. Sufficient work has been done to show that an index of vegetation cover (PD54), developed in central Australia and derived from remotely sensed (satellite) data, is applicable in India's Thar Desert. It is currently being verified with ground based data. Progress has been made with demonstrating the usefulness of India's own satellite (IRS) for monitoring vegetation cover in arid parts of India. In addition, a geographic information system (GIS) is under development for the test area.

Community entry has been achieved thus far through building on previous contacts by CAZRI scientists. Socio-economic and livestock information has been directly gathered at two villages, with the intention to survey two further villages in the next pre-winter season. CAZRI scientists have proposed a reward system to ensure future community entry, and will trial it at the next survey. Their analysis of primary data to date plus secondary data from census records has provided a detailed picture of the issues facing villagers. After the next survey period, it should be possible to develop between-village comparisons, to complement ground-based and remotely-sensed data on the extent of degradation.

Products for explaining the technical outputs of remote sensing to village people will depend on further development of the technology, in the forthcoming FY. However, experience to date suggests that, since the value of remote sensing technology lies in larger scales than individual farms, it is likely to be most useful for 'common property resources' of several hundred hectares. Hence target users will probably be regional administrators as well as individual villages.

Infrastructure and capacity building is occurring on several fronts. A remote sensing and GIS laboratory is being established and equipped at CAZRI. Indian scientists have been trained in the basic steps of interpreting and processing satellite imagery and training is continuing in image processing and interpretation. Indian scientists have also been trained in ground-based techniques for rapid collection of data and spreadsheet analysis, and have been transferring these skills to others as the project proceeds.

(iv) Likely direction of future research activities.

Development of remote sensing applications are expected to proceed as planned and collection of ground-based data should also proceed as proposed. Additional techniques for community entry will be trialled as indicated above and issues of data integration will be addressed in the forthcoming year.

Also as noted, target users now seem likely to be regional administrators as well as individual villages. Consequently, steps will be taken to engage the interest of administrators in scaling-up products to regional level.

Although there has been about a year's slippage in time, the project is now very productive. Looking further ahead, we will attempt to confirm that CAZRI has the capacity for implementation.

The project has advanced all its objectives in 2002/03. The first objective is to apply and adapt Australian-developed techniques, using satellite data, for assessing land degradation, to Indian desert environments. Testing of candidate indices of vegetation cover continued but was constrained by the failure of the 2002 monsoon and hence the lack of green vegetation to assess. 'PD54' is the preferred index, on the basis of testing so far. The raster-based Geographic Information System (GIS) in ER Mapper has been expanded to include mapped land use and land cover type, the 2002 image pair (wet and dry) and the location of roads and villages. Mapped land use and land cover type were obtained from the State Remote Sensing Advisory Centre of Rajasthan, which has offered continuing collaboration with the project. Actual and mapped intensities of land use have confirmed that grazing gradient analysis based on distance from water and villages is not feasible for detecting land degradation. Instead, we are testing spatial patterns in the extent of vegetation recovery following monsoonal rainfall for their ability to indicate land degradation. Image products have been developed and will be tested against ground data following the 2003 monsoon.

The second objective is to gather ground-based survey data on socio-economic, natural resource base and animal production for interpreting the results of remotely sensed analyses. To achieve this, the project had to first establish a trusting relationship with villagers. Initially, villagers who provided information were to be offered an honorarium to ensure continued cooperation but experience showed that this could be counterproductive for those who missed out. Instead, villagers were supplied with veterinary advice and medicines but, unfortunately, provision of medicines put local suppliers off-side. The solution was to provide prescriptions for veterinary medicines that could be supplied locally. Provision of saplings to improve village conditions and fodder supply was inappropriate because of severe drought but will occur in 2003 with a better monsoon season. Techniques for collecting and analysing vegetation data for comparison with satellite imagery had been established prior to the current reporting period. Biophysical data were collected from September to November 2002 to complement the September "end of wet season" image. No biophysical data were collected to complement the proposed purchase of an April/May 2003 image because seasonal conditions were so poor. Information on 41 farm families from Shergarh and Sointra villages had been collected prior to this reporting period. Interviews with further households in these villages plus households in Sai and Solankiya Tala, to a total of 100, were completed in 2002-03. Computer entry and analysis of these data are ongoing.

The third objective is to develop means of information exchange with village communities *and administrators* in order to explain land degradation. Following our mid-term review in February 2002, emphasis shifted from village people to include administrators as key customers for technical outputs. A workshop held at the Institute for Development Studies (IDS), Jaipur, successfully gained the interest of Government of Rajasthan officials and explored the potential for scaling-up activities to district and regional level.

The final objective is to develop capacity with Indian colleagues for their independent use of all methodologies. Most equipment has now been purchased although refurbishment of the remote sensing and GIS lab is not yet complete. Training in image processing, GIS and data analysis with Excel continued with each visit by Australian project staff to CAZRI (November 2002, April and June 2003) and is supported by written documentation.

Two key events occurred through the year. The first was the workshop at IDS, Jaipur, involving Government of Rajasthan officials, NGO representatives and IDS, CAZRI and CSIRO scientists, which established strong support for the project and for scaling up to district and regional level. Participatory processes in development planning were also recommended to accompany scaling up. The second key event was the ACIAR final review of the project, a year before completion. The review was undertaken by Dr PS Roy, of the Indian Institute of Remote Sensing, Dehradun, who offered to meet a need for formal training of Indian project staff in image processing and GIS. This was very timely in view of our lack of success in locating suitable training until now.

The project is on target to meet its objectives, particularly now that the 2003 monsoon season has been favourable.