Integrating spatially referenced social and biophysical data to explore landholder responses to dryland salinity in Australia

¹Allan Curtis, Ian Byron and Simon McDonald

¹Social Sciences Program Leader, Bureau of Rural Sciences, Agriculture, Fisheries and Forestry – Australia, PO Box E11 Kingston, ACT 2604, Australia, Phone 02 62723382, Email. Allan.Curtis@brs.gov.au

Paper accepted for publication (May 2003) in the Journal of Environmental Management

Abstract

Researchers attempting to integrate socio-economic data in watershed planning often draw on nationally collected census data. However, there are critical limitations to the usefulness of this type of data for decision makers operating at the watershed scale. In this paper we demonstrate the relevance of spatially referenced socio-economic data collected using mail surveys to random selections of rural landholders. The issue explored was dryland salinity management in two large watersheds in the Murray-Darling Basin of south-eastern Australia. Contrary to the assumptions underlying public policy in Australia, but consistent with the literature on farmer knowledge, comparisons of expert maps and landholder identified salinity sites suggested that landholders in these watersheds had excellent knowledge of the current extent of salinity on their property. Our research also suggested that salinity education was a sound investment by governments. At the same time, the expert maps failed to predict half of the saline affected sites identified by landholders. Accurately mapping the extent of salinity would seem a first step in addressing this nationally significant land degradation issue.

Key words

Dryland salinity, Australia, watershed management, integration.

Introduction

In this paper we discuss research in two large watersheds in the Murray-Darling Basin (MDB) of south-eastern Australia. Our research explored the assumptions that private landholders were often unaware of insidious issues such as dryland salinity (Vanclay 1992); and that low levels of awareness and concern were important reasons for limited action by private landholders (MDBC 1990). This research examining landholder awareness of dryland salinity, the veracity of expert maps, and the efficacy of community education programs forms part of the wider international literature on farmer knowledge (Kloppenburg 1991; Thompson and Scoones 1994) and adoption (Roling 1988; Vanclay 1992).

Most researchers attempting to integrate socio-economic data in watershed planning have drawn on nationally collected census data (Radeloff et al. 2000). There are critical limitations to the usefulness of census data, usually available at the local government scale, for decision makers wanting to understand and change individual landholder behaviour (Shultz et al. 1998). In our MDB case studies, we integrated expert data about saline affected areas and information provided by landholders through a mail survey process. The principal purpose of this paper is to demonstrate the usefulness of this spatially referenced socio-economic data collected from individual landholders.

Background

The management of dryland salinity is now recognised as a major challenge facing Australia, particularly in the MDB, often referred to as Australia’s food bowl. On a scale that is comparable with the Colorado River Basin, the MDB embraces substantial areas of the states of New South Wales, the Australian Capital Territory (ACT), Queensland, Victoria and South Australia. Research suggests that more land will be affected by dryland salinity in the near future, placing even greater pressure on water quality, agricultural production, biodiversity, infrastructure and cultural heritage (MDBC 1999).

Although the Commonwealth (federal) government has greater financial resources, natural resource management authority rests primarily with the governments of the six Australian states and two territories (ACT and the Northern Territory). By 1992, most Australian states had established regional Catchment Management Committees (CMC). CMC members are mostly Ministerial appointees, frequently including a mix of regional community representatives and agency staffers. CMC are responsible for developing and implementing regional catchment strategies that guide the expenditure of state and Commonwealth natural resource management funds (Curtis and Lockwood 2000; Ewing 2000).

As part of a national response to dryland salinity, the Commonwealth and State governments recently committed 1.4 billion dollars Australian over seven years through a National Action Plan for Salinity and Water Quality (NAP). Most of the NAP investment decisions will be devolved to the regional CMC. Prior to the NAP there had been large investments in natural resource management programs by Commonwealth and State governments through the National Landcare Program and subsequently, the Natural Heritage Trust (Curtis and Lockwood 2000). Again, dryland salinity was a priority issue.

A key assumption underpinning these programs is that landholders are frequently unaware of the extent and impact of less obvious forms of land degradation, such as dryland salinity and soil acidity (MDBC 1990; ASCC 1991). For much of the past twenty years, community education activities to raise awareness and understanding of issues have been an important element of these programs.

This research was a collaborative effort between the authors, staff from the State Department of Natural Resources and Environment (DNRE), the two regional Catchment Management Authorities (CMC) and private landholders. These partnerships facilitated the integration of socio-economic data collected using mail surveys to private rural landholders and other data layers held within DNRE and by DNRE partners, including the large consulting firm, Sinclair Knight Merz (SKM). The SKM data layers had just been compiled as part of the work SKM completed for the Murray-Darling Basin Commission’s seminal salinity audit (MDBC 1999).

Situation

Our discussion in this paper draws on research in two adjacent watersheds in the North East region of the state of Victoria:

Goulburn Broken Watershed (GBW); and
Ovens Watershed (OW). [Figure 1]

The GBW covers 2.3 million hectares (17 per cent of the state of Victoria), including 1.9 million hectares of non-irrigated land that is referred to as the Goulburn Broken Dryland (GBD). The research reported in this paper for the GBW was undertaken in the GBD. The OW covers a smaller area of 780,000 hectares of largely non-irrigated land.

Both watersheds begin on the continental side of the Australian Alps at elevations around 2000 metres ASL, and fall through a series of river valleys towards the westerly flowing Murray River at around 200 metres ASL. The climate is typical of south-eastern Australia, with hot, dry summers and cool, wet winters, but there are significant variations with altitude, in that temperatures are lower and rainfall higher in the mountains. Much of the mountains and foothills are covered in eucalypt forests and most of these areas are publicly owned. European systems of agriculture on privately owned land dominate the lower elevations and this land has little remaining native vegetation. There are important areas of irrigated agriculture, mostly horticulture and dairying, but dryland farming for sheep and cattle grazing and broad acre cropping (canola, wheat) occupy most cleared land. Primary production and associated processing industries, and tourism are the main contributors to economic wealth. Agricultural production from these watersheds contributes more that 25 per cent of Victoria’s total export income.

In both watersheds there are numerous small towns (population <3,000) and at least one regional city (Wangaratta with 15,000 and Shepparton with 35,000). There has been considerable rural subdivision (Melbourne has a population of 3.5 millions and is between two and four hours away by car) and substantial proportions of properties in both watersheds are operated as lifestyle as opposed to farming enterprises. Both the GBW and OW support major agricultural industries, food processing, forestry (including pine plantations) and tourism activities. Soil erosion, water logging and dryland salinity have been some of the unintended consequence of land clearing in both watersheds.

Figure 1 about here

The GBW and OW represented an ideal setting for research exploring awareness and concern about dryland salinity. This is a priority issue in both watersheds and as adjacent watersheds, they share similar environmental, socio-economic and political contexts. At the same time, there are important contextual differences. Compared to the GBW, dryland salinity in the OW emerged later (post 1990) and is expected to have less severe off-site impacts. The GBW has been identified as the largest contributor of saline discharges into the River Murray (MDBC 1999, p.20). There has been a substantial investment in the OW for research, community education (for example, field days, farm walks, monitoring programs, schools education, property planning) and on-ground work to address dryland salinity. However, this investment occurred much later and on a much smaller scale than was the case in the GBW. Indeed, the GBW has been the focus of one of the most sophisticated and concerted efforts by government agencies, community organisations and landholders to address dryland salinity in Australia. In recent years, around 60 per cent of federal and state funds expended on natural resource management in Victoria have been invested in the GBW (Curtis and Lockwood 2000). Much of this investment, including at least $20 million over the past 15 years, has been focussed on dryland salinity.

Current recommended practices (CRP) expected to improve the management of dryland salinity by private landholders in the GBW and OW include; the revegetation of recharge and discharge zones; adoption of farm forestry; replacing annual pastures with perennial pastures (mostly introduced); and intensifying grazing to increase water use (GBCLPB 1996; OBWQWG 2000).

Problem and purpose

Researchers have established a significant association between increased awareness of land degradation issues and landholder adoption of practices likely to ameliorate those issues (Vanclay 1992; Curtis and De Lacy 1996). It has also been established that most landholders are more concerned about off-property and district impacts of land degradation compared to on-property impacts (Vanclay 1992; Curtis et al. 2002). These findings have been interpreted as evidence of landholder denial of on-property problems and this was seen as an important explanation for the reluctance of landholders to respond to dryland salinity (Vanclay 1992).

Attempts to integrate socio-economic and biophysical data to inform watershed management usually rely on census data. This type of analysis can yield useful information (Radeloff et al. 2000), but is unlikely to contribute to improved understanding of landholder behaviour, including assessments of landholder awareness of natural resource issues, that is required by watershed management decision makers. In the first instance, it is difficult for state or national data collection processes, such as the household and farm census, to address most of the topics for which data is needed. Curtis et al. (2001) found that information from landholders about their awareness and concern about issues, values attached to their land or property, knowledge of processes or management practices, confidence in recommended practices, and long-term plans was critical for effective watershed management. Secondly, information that is collected by existing processes such as the household and farm census is usually only available at a local government scale and has limited application when the task is to unravel decision making by individual landholders (Schultz et al. 1998). A final limitation is that census data is collected intermittently, every five years in Australia, and with the time lag between collection and release of data, available information can be up to eight years out of-date. For example, in Australia, the most recent census were in 1996 and 2001. However, the Australian government is still using 1996 data for some topics where the Australian Bureau of Statistics has not released 2001 data.

Despite the limitations of census data, there have been few attempts to directly capture and integrate spatially referenced socio-economic data to address watershed management issues. There are important difficulties related to the high costs of data collection (Schultz et al. 1998); difficulties inherent in building the partnerships required to access different data layers, including legitimate concerns about the loss of intellectual property rights and breaching the confidentiality of personal information. There are also concerns that alternate data collection instruments, such as mail surveys to rural landholders, can achieve response rates that will engender confidence that data collected is representative (Curtis et al. 2001).

There is some discussion in this paper of the broader aspects of landholder adoption of current recommended practices (CRP) to address dryland salinity. However, the focus of the paper is on the outcomes of efforts to integrate spatially referenced socio-economic and biophysical data layers. The key research questions were:

What proportion of landholders thought they had areas where vegetation was affected by dryland salinity?
What proportion of landholders was unaware that they had areas affected by dryland salinity according to the expert maps (discharge sites and depth to saline groundwater)?
How successful were the expert maps in predicting the areas affected by dryland salinity that were identified by the landholders?
Was awareness of saline affected areas or the extent of concern about dryland salinity linked to adoption of CRP?

The experience gained through our research in these two watersheds therefore has wider relevance for social research and watershed management. Indeed, the Commonwealth and State governments in Australia are drawing on this experience to undertake a national watershed project coordinated by the Bureau of Rural Sciences, a Commonwealth government research organisation.

Methods

The mail survey processes

This research used a mail survey to collect information from rural landholders in the GBW and OW watersheds. The GBW survey was completed during 1999 and the OW survey in 2001. There was therefore the opportunity to learn from experience and refine the research methodology. At the same time, some caution needs to be exercised when making direct comparisons between data from the two watersheds, particularly information about on-property profitability that could be expected to have improved since 1999 as cattle prices, and more recently wool prices have firmed.

As part of the process of identifying variables to be included in the mailed survey, the research team examined regional and state planning documents, held discussions with industry partners and other key stakeholders and conducted a search of literature on adoption of CRP for dryland salinity. We also drew on the experience of the research team with farm forestry, introduced and native perennial grasses and revegetation with native species (Curtis and DeLacy 1996; Curtis and Race 1996; Millar and Curtis 1997).

The main topics included in the 12-page survey booklets relevant to this paper included.

Importance of issues affecting property and district.
Self-assessment of knowledge for different topics.
Awareness of on-property salinity. Survey respondents were asked two questions:

Was there any part of their property where vegetation was showing signs of the effects of salinity?
What was the area of land where vegetation was showing signs of the effects of salinity?

Adoption of current recommended practices (CRP).
Background socio-economic data.

In early 1999 when the GBW study commenced we contacted the eight local government councils and asked them to assist with the compilation of our mailing list. This process met with limited success. At this time, local governments in Victoria were being amalgamated and some of the new entities were not able to provide revised ratepayer lists. The next best option was to use Country Fire Authority (CFA) rural property maps to identify rural landholders and then use Commonwealth electoral rolls and phone books to source their postal addresses. The major limitation with the CFA maps was that they were a little out-of-date. Using this process, a total of 6,449 rural properties were identified for the dryland areas (GBD) in the GBW. These property listings were entered into a Microsoft Excel spreadsheet and a random sample of 1,640 properties (approximately 25 per cent) was generated.

By the time we commenced the OW study in early 2001 the local government amalgamations of 1999 had been bedded down. The OW covered three local government councils and their staff agreed to provide access to their ratepayer data bases that included a property identification field for most properties. This field was then matched to cadastral boundaries in Arcview GIS (Geographic Information System). The centroids of these cadastral boundaries were calculated and were then representative of the location of each landholder. A total of 8,658 rural properties were identified in the OW. From this list a random sample of 1,000 properties (approximately 12 per cent) was generated with matching postal addresses.

The survey design and mail-out processes followed Dillman’s (1979) Total Design Method. Pre-testing of the survey instrument was undertaken using a number of small group workshops with landholders who had been identified by agency and industry partners. The initial survey mail out was followed by up to three reminder/thank you cards in successive weeks. A second mail out, with one reminder/thank you card, was undertaken to all non-respondents to the first mail out.

Surveys that were returned to sender or sent back due to the landholder no longer residing at the property were removed from the original sample. Cases where landholders that were too old, ill, deceased, the property was a residential lot (considered by the agency staff involved to be <4ha in the GBW and <10ha in the OW) or the property had been sold were also removed from the original sample. In the GBW study, a final sample of 1,021 was left. With 480 completed surveys returned, the overall response rate was 47 per cent. In the OW, a final sample of 854 was left. With 568 completed surveys returned, the final response rate was 67 per cent.

Data analysis

Findings presented in this paper are derived from analyses undertaken using a range of descriptive statistics and binary logistic regression. All analyses were undertaken using the SPSS statistical package.

Mail survey data were also entered into an Arcview GIS that contained other data layers, including salinity discharge sites provided by DNRE (Allan et al. 1997; CLPR 2000), and depth to water table and ground water salinity levels (SKM 2000). The discharge maps (Allan et al. 1997; CLPR 2000), had been compiled over time employing a mix of air photo interpretation and intensive field inspections by agency staff and consultants. At this point it should be noted that there was a two to three year time lag between the preparation of the discharge maps in each watershed and the mail out of the landholder survey. It was expected that the discharge maps would slightly under-estimate the extent of saline affected vegetation in both watersheds. The depth to water table and ground water salinity level data layers (SKM 2000) had been prepared using interpretations of time-series data from a network of bores and other hydrogeological information. These data layers had been prepared at much the same time as the landholder survey and were assumed to be up-to-date. It was beyond the scope of this research to test the validity or reliability of the data layers provided by SKM or DNRE.

The expert and mail survey data layers were then used in analyses comparing landholder identified dryland salinity affected areas with salinity mapping by experts. This assessment involved comparing landholder proximity (using a spatial intersection) to two types of salinity prone areas.

a 1 kilometre buffer of the discharge layer (for both the GBW and OW); and
an intersection using the Geoprocessing Wizard of areas with a level of total dissolved salt of 3501 mg/L and areas with a water table less than 2 metres from the surface (data layers only available for GBW).

A 1 kilometre buffer was adopted to provide some margin of error when comparing the location of discharge sites mapped on a 1:25,000 sheet with landholder reported saline affected sites that could only be mapped as a property location. Two metres is the widely accepted threshold for saline ground water to affect surface vegetation (MDBC 1999, p.2). Salinity tolerance varies with vegetation type, however the MDBC (1999) study accepted that ground water with a total dissolved salt content higher than 3501 mg/L will affect most perennial vegetation on farms (Hoxley pers. comm.).

Discussion of findings

Dryland salinity reported by landholders

In the GBW study, 13 per cent of the respondents (N=456; n=61) said they had areas on their property where vegetation showed signs of the effects of salinity. For most respondents, the area affected was relatively small (median of 4 hectares). A similar picture emerged in the OW study, with 11 per cent of respondents (N=550; n=62) reporting they had areas of vegetation on their property that showed the effects of salinity [Table 1]. Widespread clearing of native vegetation occurred pre-1900 in the GBW, much earlier than in the OW and it was expected that dryland salinity would affect a larger proportion of land in the GBW. This hypothesis was confirmed by the finding that the area of saline affected land reported by landholders was smaller in the OW (median of 2 hectares). However, less than one per cent of the area surveyed was reported as being affected by dryland salinity in either watershed.

Table 1: Comparing landholder perceptions of salinity with expert maps
Goulburn Broken Watershed 1999 (N=480), Ovens Watershed 2001 (N=568)

Topic	GBW	OW
Number of respondents to question about salinity	456	550
% who reported no saline affected areas on property	87%	89%
% who reported a salt problem	13%	11%
% who reported no salinity but expert maps differed	6%	10%
% who reported a salt problem but expert maps didn’t show a saline affected area	51%	61%
% who said no salinity but expert maps predict a salinity problem in the future	19%	N/A

What proportion of landholders was unaware that they had areas affected by dryland salinity according to the expert maps?

Analyses using maps of salinity discharge sites provided by DNRE (Allan et al. 1997; CLPR 2000), and depth to ground water and ground water salinity (SKM 2000) suggested that respondents had a very high level of awareness and preparedness to acknowledge current, visible, dryland salinity impacts.

Five per cent (N=395; n=18) of those in the GBW and 10 per cent (N=481; n=47) of those in the OW who reported no areas with vegetation that showed the effects of salinity were within one kilometre of a discharge site identified on DNRE maps.
In the GBW, only six (1.52 per cent) of those who reported no areas with vegetation that showed the effects of salinity had ground water within two metres of the surface and a total dissolved salt content higher than 3501 mg/L as mapped by SKM (SKM 2000).

Overall, the expert maps appeared to contradict the claims by six per cent (n=24) of GBW respondents and 10 per cent (n=47) of OW respondents who said they had no areas where vegetation showed the effects of salinity [Figure 2]. Assuming the expert maps are correct, that is they didn’t incorrectly predict salinity, these respondents were “unaware” they had saline affected areas. In other words, almost 90 per cent of the respondents who said they had no areas currently affected by salinity were correct according to the expert maps. It seems that landholders have a very accurate level of awareness of the areas currently affected by dryland salinity. This finding is contrary to earlier research suggesting farmers are not aware of salinity and/or are not prepared to acknowledge the extent of salinity problems (Vanclay 1992).

Figure 2 about here

We also explored the extent that there were differences between those who were aware that they had saline affected areas and the “unaware” group identified above. These analyses suggested that those respondents who were aware of salinity affected areas were significantly more likely to operate larger properties, work more hours on-property, be Landcare members (local or community-based watershed organisations) and report work funded by government [Table 2].

Table 2: Comparing those who reported dryland salinity with those the expert maps suggested were unaware they had saline affected areas
Goulburn Broken Watershed 1999 (N=480), Ovens Watershed 2001 (N=568)


Social and farming variables	Ovens Watershed			Goulburn Broken Watershed
Social and farming variables	Aware	Unaware	Sig.*	Aware	Unaware	Sig.*
Property size (median)	258 ha	132 ha	p<0.05	390 ha	51 ha	p<0.001
Hours worked on-property (median)	No significant difference			45 hours	25 hours	p<0.05
% Landcare members	75%	51%	p<0.05	No significant difference
Involved in government funded program	57%	26%	p<0.001	No significant difference

*Significant using Kruskal-Wallis Chi-square test.

The lack of awareness displayed by a small number of landholders (24 in the GBW and 47 in the OW) may be explained by the possibilities listed below.

Those with lower awareness spend fewer hours in on-property management and may not have identified saline affected areas.
Some landholders cannot identify the subtle changes in vegetation that indicate dryland salinity.
People may be in denial and do not want to acknowledge they had saline affected areas.
The expert maps have falsely predicted saline affected areas.

There has been a large, almost unprecedented level of investment over more than a decade in community education about dryland salinity in the GBW. Landcare groups have operated for at least 15 years in the OW (Curtis and De Lacy 1996) and there has been a strong salinity education focus over the past five years through these groups and government funded extension programs. It seems that the investment in community education in both watersheds has contributed to the high levels of awareness of dryland salinity identified in this paper.

Did the expert maps predict landholder identified dryland salinity?

There was also the opportunity to examine the efficacy of the expert maps by assessing their capacity to predict areas affected by salinity as identified by landholders [Figure 3]. With the two to three year time lag between the preparation of the discharge maps and the landholder survey it was expected that the expert maps would slightly under-estimate saline affected areas. This issue was not expected to be as important in the GBW where there was the more up-to-date SKM data layers.

The expert maps correctly predicted areas where salinity was affecting vegetation for 49 per cent (N=61; n=30) of the GBW and 39 per cent (N=62, n=24) of the OW respondents that reported saline affected areas. This research suggests that the expert maps had failed to predict between 50 and 60 per cent of the areas affected by salinity in the two watersheds.

It is unlikely that landholders would deliberately overstate the extent of salinity on their property. However, there is a possibility that some landholders have failed to distinguish between water logged and saline affected areas. As Clark (2000) explained, some salt-tolerant species grow in both saline and non-saline conditions. At the very least, the finding that the expert maps failed to predict a large proportion of the saline affected areas identified by landholders warrants further survey work. The budget and scope of the two projects reported in this paper did not permit this work to be undertaken.

Most of the landholder identified saline sites that were not predicted by the expert maps (GBW, N=31; OW, N= 38) are located south east of the Hume Freeway in the foothills of the Great Dividing Range (GBW, n= 20; OW, n=30) [Figure 1 and Figure 2]. This was a statistically significant pattern in the OW (Chi-square = 18.635 p<0.001) and a trend in the GBW (Chi-square = 3.503 p=0.061). Discussions with our agency partners revealed that salinity problems emerged earlier on the plains to the north west of the Hume Freeway and that this is the area where most resources have been invested in field studies to identify saline sites.

Figure 3 about here

Concern about salinity

In the GBW study, respondents expressed low levels of concern about a range of potential economic, environmental and social impacts of dryland salinity (using a five point response option, mean scores for each topic were below 2.6 out of a possible five). For example, only 32 per cent of respondents said they were ‘alarmed’, ‘very concerned’ or ‘concerned’ about the potential threat posed by rising water tables on the long-term productive capacity of their property [Table 3].

In the OW study, respondents were asked to provide an assessment of the importance of a range of issues, including dryland salinity. Respondents were asked to indicate the importance of dryland salinity as a threat to the quality of river water in their district, the long-term productive capacity of land in their district and the long-term productive capacity of their property. Dryland salinity was not rated highly as an important issue in the OW [Table 3], with mean scores for the three topics below 2.8 out of a possible 5. The highest rating salinity statement was ranked nine out of 16 statements.

Table 3: Concern about the impact of dryland salinity
Goulburn Broken Watershed 1999 (N=480), Ovens Watershed 2001 (N=568)


Potential impacts of rising water tables	n	Goulburn Broken Watershed					Mean score**
Potential impacts of rising water tables	n	Alarmed	Very concerned	Medium concern	Small concern	Not a problem	Mean score**
Threat to the long-term productive capacity of this area.	444	7%	21%	21%	22%	31%	2.51
Polluting fresh ground water in this area.	442	10%	18%	17%	19%	37%	2.44
Threat to the long-term viability of the local economy.	442	5%	19%	20%	24%	32%	2.41
Contributing to the decline of habitat or wildlife in this area.	443	6%	15%	19%	24%	36%	2.32
Detracting from attractiveness of area as a place to live.	443	3%	12%	21%	25%	39%	2.16
Reducing the value of my property.	443	5%	12%	16%	24%	43%	2.12
Threat to long-term productive capacity of my property.	442	4%	14%	14%	23%	45%	2.09
Perceived threat of salinity	n	Ovens Watershed					Mean score
Perceived threat of salinity	n	Very important	Important	Some	Minimal	Not important	Mean score
Threat to long-term productive capacity of land in this district.	538	18%	33%	20%	14%	15%	2.75
Threat to the quality of river water in this district.	534	25%	36%	16%	12%	11%	2.48
Threat to long-term productive capacity of my property.	535	38%	33%	13%	10%	6%	2.14

** Score where 1 = not a problem/not important through to 5 = alarmed/very important

It seems that low levels of concern in the GBW and OW about the impacts of dryland salinity reflect the current restricted extent of visible dryland salinity (MDBC 1999). Most respondents were not experiencing visible salinity problems and the majority of those that were experiencing problems, had only small areas where salinity had affected vegetation. It seems that many respondents believe they can “live with salt”.

While respondents have very good awareness of current saline affected areas they may not be aware of the predicted increase in the area affected by dryland salinity. For example, the most recent assessments predict that over the next 20 years the area affected by salinity in the GBW will increase by 160 per cent (MDBC 1999, p.30). Unfortunately, there was no way for us to assess whether individuals who said that they don’t have saline affected areas can be expected to have a problem in the future. The MDBC (1999) had predicted the total area of land in GBW that was likely to be affected by dryland salinity in the future using bore data that monitors rising water tables. This approach meant that it was not possible to identify the specific parts of the landscape, and therefore individual properties, where salinity was expected to impact in the future.

For the GBW we were able to assess the extent that landholders currently without saline affected areas are located in areas that have saline ground water. Assuming that in every case this ground water would rise to within two metres of the surface, an additional 76 GBW respondents would be affected by dryland salinity. Combining this group with those already affected by salinity (13 per cent) would mean that 30 per cent of all respondents (137 of 456) would be affected. Again, most respondents will not be affected by dryland salinity.

Adoption of recommended practices

Findings from the two watersheds suggested that awareness and concern about dryland salinity were linked to adoption of CRP thought to enhance the management of dryland salinity.

Using binary logistic regression, OW landholders who reported saline affected areas were estimated as being 1.7 times more likely to establish perennial pastures (Wald = 4.063, p = 0.044, Exp(B) = 1.695) and 3.3 times more likely to plant trees (Wald = 20.637, p < 0.001, Exp(B) = 3.346) than those who said they had no saline affected areas. In the GBW, landholders who reported saline affected areas were 2.6 times more likely to establish perennial pastures (Wald = 7.581, p = 0.006, Exp(B) = 2.595) than those who said they had no saline affected areas.

We also explored the extent of differences in adoption of CRP between those who were aware that they had saline affected areas and the “unaware” group (expert maps suggested they were wrong in saying they had no saline affected areas) identified earlier. Using logistic regression, OW landholders who reported saline affected areas were estimated to be 3.4 times more likely to plant trees than those who were “unaware” they had saline affected areas (Wald = 8.811, p = 0.003, Exp(B) = 3.394). In the GBW, landholders who reported saline affected areas were estimated to be 4.1 times more likely to establish perennial pastures (Wald = 6.669, p = 0.010, Exp(B) = 4.132).

In the OW, higher concern about the impacts of dryland salinity (as measured by a scale comprising the three items in Table 2) was also linked to significantly higher adoption of trees planted (Wald = 9.585, p = 0.002, Exp(B) = 1.036).

Conclusions

This paper presented one of the few examples where researchers have moved beyond the use of census data when integrating spatially referenced socio-economic and biophysical data to address watershed management issues. In this case the focus was on exploring landholder awareness of dryland salinity, the veracity of expert salinity maps, and the efficacy of government responses to dryland salinity, including through community education programs. While the research findings contribute to the wider literature on farmer knowledge and adoption, the most important outcome was that the research demonstrated the usefulness of integrating spatially referenced socio-economic data collected by surveying landholders.

It had been assumed that landholders were either unaware of the extent and impact of less obvious forms of land degradation, such as dryland salinity, or were in a state of denial. Comparisons of landholder identified salinity affected areas and those predicted by expert maps suggested that landholders in the two watersheds in this study had excellent awareness of the current extent of salinity on their properties.

A very small proportion of landholders was identified as being unaware of saline affected areas on their property. Compared to this unaware group, those who reported saline affected areas on their property were significantly more likely to be members of Landcare, be involved in government programs, operate larger properties and work longer hours on-property. At the same time, those who acknowledged they had saline affected areas had adopted CRP for salinity mitigation at significantly higher levels than those who were thought to be unaware of salinity affected areas on their property. Those who reported saline affected areas also had higher adoption of CRP than all of the respondents who said they did not have saline affected areas. These findings suggested that awareness is linked to adoption and that the substantial investment in community education in these regions had been successful in raising salinity awareness and had contributed to the adoption of CRP.

By contrast, the expert maps (discharge sites and depth to saline ground water) failed to predict saline affected sites identified by half the respondent landholders in both studies. The time lag between the compilation of the expert maps and the landholder survey explains some of this difference. Nevertheless, the finding that the expert maps prepared for the GBW - a watershed where there has been a sustained attempt to map ground water and discharge sites - under-stated salinity affected areas by 50 per cent is a cause for concern. Given the importance of dryland salinity and the upcoming 1.4 billion dollar investment by Australian governments under the National Action Plan for Salinity and Water Quality, it is critical that regional natural resource management planning bodies (CMC) can accurately map the extent of dryland salinity.

Most survey respondents did not report visible signs of dryland salinity on their property and our calculations for the GBW suggested that most of these landholders are unlikely to experience dryland salinity in the future. Given these findings, it was not surprising that most respondents were unconcerned about the impacts of dryland salinity, appear to believe they can, or must, “live with salt” and are very reluctant to invest in recommended practices for salinity amelioration. This information suggests that efforts to address natural resource management issues should not rely on appeals to respond to the threat of salinity and must address the range of values landholders attach to their properties. If salt originating in upland watersheds like the GBW and OW is a critical issue for others, including governments, this needs to be acknowledged and addressed through cost-sharing with downstream landholders; supporting landholders to move into profitable emerging enterprises; and government funding for natural resource management (Curtis and Lockwood, 2000).

With refinement, the mail survey process produced a high response rate (67 per cent in the OW) and was able to gather information at the property scale not available from other sources, including the national census of households and farms. By building strong collaborative partnerships with government agencies and consultants, the researchers were able to secure access to vital data layers and successfully overcome issues related to intellectual property rights and privacy.

Acknowledgments

Project funds and in-kind contributions were made by the Murray-Darling Basin Commission, the Goulburn Broken and North East Catchment Management Authorities and the Department of Natural Resources and Environment (DNRE). Mark Cotter and Royce Sample were the project managers for these organisations. Staff from Sinclair Knight Merz (SKM) provided access to SKM spatial data and important advice about how to interpret that information. Michael Lockwood, Marike Van Nouhuys, Megan Graham and Wayne Robinson from Charles Sturt University made important contributions to aspects of project development, data management and analysis and report preparation. Neil Barr (DNRE) and Mike Read (Read Sturgess & Associates), who were working on related research projects, provided valuable assistance. The authors particularly thank those landholders that contributed to pre-testing of the survey instrument, completed surveys and were members of the expert panel that reviewed preliminary research findings in the GBW.

References

Allan, M., Reynard, K. and Williams, S. (1997). Dryland salinity discharge sites (linear and point), Department of Natural Resources and Environment, Melbourne, Australia.

ASCC (Australian Soil Conservation Council). (1991). Decade of Landcare plan, ASCC, Canberra, Australia.

Clark, R.M. (2000). Establishing three discharge monitoring sites in the North East salinity region, Centre for Land Protection Research, Monitoring Report No. 23, Department of Natural Resources and Environment, Melbourne.

CLPR (Centre for Land Protection Research). (2000). Discharge sites (GIS map layers), CLPR, Bendigo, Victoria.

Curtis, A. and De Lacy, T. (1996). Landcare in Australia: does it make a difference. Journal of Environmental Management 46, 119-137.

Curtis, A. and Lockwood, M. (2000). Landcare and catchment management in Australia: lessons for state-sponsored community participation. Society and Natural Resources 13, 61-73.

Curtis, A. and Race, D. (1996). Review of socio-economic factors affecting regional farm forestry development in Australia. Report to Rural Industries Research and Development Corporation. Albury, New South Wales: The Johnstone Centre, Charles Sturt University.

Curtis, A., Lockwood, M. and MacKay, J. (2001). Exploring landholder willingness and capacity to manage dryland salinity in the Goulburn Broken Catchment. Australian Journal of Environmental Management 8, 20-31.

Curtis, A., Graham, M., Byron, I., Lockwood, M. and McDonald, S. (2002). Providing the knowledge base to achieve landscape change in the Ovens catchment. Albury, New South Wales: The Johnstone Centre, Charles Sturt University.

Dillman, D.A. (1979). Mail and telephone surveys: the total design method. New York: Wiley.

Ewing, S. (2000). The place of Landcare in Catchment Management Structures. In Proceedings of the International Landcare 2000 Conference: Changing Landscapes ~ Shaping Futures, 2-5 March 2000. Melbourne, Australia.

GBWLPB (Goulburn Broken Catchment and Land Protection Board). (1996). Our Catchment & Community-The next step: Goulburn Broken Dryland Salinity Management Plan (Five-Year Review 1990-1995 and Future Direction). Shepparton, Victoria: GBWLPB.

GBWMA (Goulburn Broken Catchment Management Authority). (1998). Catchment Strategy Final Draft. Shepparton, Victoria: GBWMA.

Kloppenburg, J. (1991). Social theory and the de/reconstruction of agricultural science: local knowledge for an alternative agriculture. Rural Sociology 56, 519-548.

Millar, J. and Curtis, A. (1997). Perennial Grasses: finding the balance. Australian Journal of Soil and Water Conservation 10 (1), 21-28.

MDBC (Murray-Darling Basin Commission). (1990). Natural Resources Management Strategy. Canberra: MDBC.

MDBC (Murray-Darling Basin Commission). (1999). The Salinity Audit of the Murray-Darling Basin. Canberra: MDBC.

Ovens Basin Water Quality Working Group. (2000). Ovens Basin Water Quality Strategy, An action plan within the North East regional catchment strategy. Wangaratta, Victoria: North East Catchment Management Authority.

Radeloff, V.C., Hagen, A.E., Voss, P.R., Field, D.R. and Mladenoff, D.J. (2000). Exploring the spatial relationship between census and land-cover data. Society and Natural Resources 13, 599-609.

Roling, N. (1988). Extension science: Information systems in agricultural development. Cambridge: Cambridge University Press.

Schultz, S., Saenz, F. and Hyman, G. (1998). Linking people to watershed protection planning with a GIS: A case study of a central American watershed. Society and Natural Resources 11, 663-675.

Scoones, I. and Thompson, J. (eds.) (1994). Beyond farmer first. Rural people’s knowledge, agricultural research and extension practice. London: Intermediate Technology Publications.

SKM (Sinclair Knight Merz). (2000). Watertable salinity information, and Broken/Goulburn depth to watertable information (GIS map layers). Armadale, Melbourne, Victoria: SKM.

Thompson, J. and Scoones, I. (1994). Challenging the populist perspective: rural people’s knowledge, agricultural research and extension practice. Agriculture and Human Values 11 (2/3), 58-76.

Vanclay, F. (1992). The social context of farmers’ adoption of environmentally-sound farming practices. In Agriculture, environment and society (G. Lawrence, F. Vanclay, and Furze, B. eds). Melbourne: Macmillan.

Water