No description available.

No description available.

In much of Australia and the United States, the almost universal right to have a house on rural allotments has resulted in rural living settlement in areas intended, by land use planning policy, to be used for agriculture. Such ‘unplanned’ rural living has a range of potential adverse impacts, including, among other things, loss of agricultural production, land use conflicts with agriculture, land value constraints on farm restructuring, and changes to the rural landscape character. Given the strongly held community perception of a house as a development right, this research was motivated by the need to understand the effects of that right, in order to assess the potential value of any change in policy. An integral consideration was the influence of the size of allotments with any such right. The research was also driven by an interest in explaining the pattern of unplanned rural living settlement, i.e. why are particular allotments converted and others not? Largely in the absence of previous studies of unplanned rural living as a distinct phenomenon, there was scope for this thesis to make a contribution to knowledge in these respects. The research involved three stages of empirical analysis. Firstly, a Queensland-wide survey of land use conversion from sugarcane production supported the choice of the Bundaberg region as the study area. Contingency table analyses were then conducted into the current land use and other attributes of land in the study area that, in 1980, was used for sugarcane. Those analyses identified distinguishing attributes of the different land use sub-sets of allotments, and informed the selection of observation units for the third stage. The third stage provided the principal basis for fulfilling the research objectives. It focused on logit models of the choice, at the time of sale, between using a particular allotment for unplanned rural living or sugarcane production. Those analyses provided an indication of the role of individual land attributes while controlling for the effects of other attributes. It was found there had been limited conversion of suitable, productive sugarcane land to unplanned rural living. Compared to those allotments that remained in sugarcane production, allotments converted to rural living were, on average, much smaller, of less value, had lower agricultural production potential, and were situated in more undulating and forested landscapes further from Bundaberg. These clear distinctions between rural living and sugarcane allotments contributed to logit models with high explanatory power. The expected productive income of allotments had overwhelming weight and was highly statistically significant in explaining the land use choice between unplanned rural living and sugarcane. This is a notable finding, because some other studies have not shown such a strong relationship between land use and measures such as soil productivity. There was a degree of natural coincidence between more undulating and forested landscapes and lower productive potential. Notably, landscape attributes appeared to be of secondary importance in determining land use. The presence of a house at the time of sale was not a statistically significant explanatory factor for land use. This finding, together with that of limited conversion of suitable, productive land to unplanned rural living, meant there was a lack of evidence to support a change to the right to have a house on rural allotments. The findings did indicate that the allotment area required to avoid conversion to unplanned rural living was different to the area required for agricultural viability. In the circumstances of this study area and period, 20 ha would have been an acceptable minimum to retain suitable cropping land in productive use. This is compared to the 60 ha suggested as necessary for farm viability. However, the potential for changed economic conditions to reduce the productive value of agricultural land, and increase demand for rural living, makes it appropriate to be cautious about permitted allotment sizes. A detailed conceptual framework informed the selection of observation units and variables for the third stage of analyses. The framework helped to confirm allotment sales as the observation units, and provided the rationale for excluding the difficult to measure influences of a landowner’s attachment to the land, and inertia, as explanatory variables. The high explanatory power of the logit models provides support for the approach, which may be useful in future studies.

In much of Australia and the United States, the almost universal right to have a house on rural allotments has resulted in rural living settlement in areas intended, by land use planning policy, to be used for agriculture. Such ‘unplanned’ rural living has a range of potential adverse impacts, including, among other things, loss of agricultural production, land use conflicts with agriculture, land value constraints on farm restructuring, and changes to the rural landscape character. Given the strongly held community perception of a house as a development right, this research was motivated by the need to understand the effects of that right, in order to assess the potential value of any change in policy. An integral consideration was the influence of the size of allotments with any such right. The research was also driven by an interest in explaining the pattern of unplanned rural living settlement, i.e. why are particular allotments converted and others not? Largely in the absence of previous studies of unplanned rural living as a distinct phenomenon, there was scope for this thesis to make a contribution to knowledge in these respects. The research involved three stages of empirical analysis. Firstly, a Queensland-wide survey of land use conversion from sugarcane production supported the choice of the Bundaberg region as the study area. Contingency table analyses were then conducted into the current land use and other attributes of land in the study area that, in 1980, was used for sugarcane. Those analyses identified distinguishing attributes of the different land use sub-sets of allotments, and informed the selection of observation units for the third stage. The third stage provided the principal basis for fulfilling the research objectives. It focused on logit models of the choice, at the time of sale, between using a particular allotment for unplanned rural living or sugarcane production. Those analyses provided an indication of the role of individual land attributes while controlling for the effects of other attributes. It was found there had been limited conversion of suitable, productive sugarcane land to unplanned rural living. Compared to those allotments that remained in sugarcane production, allotments converted to rural living were, on average, much smaller, of less value, had lower agricultural production potential, and were situated in more undulating and forested landscapes further from Bundaberg. These clear distinctions between rural living and sugarcane allotments contributed to logit models with high explanatory power. The expected productive income of allotments had overwhelming weight and was highly statistically significant in explaining the land use choice between unplanned rural living and sugarcane. This is a notable finding, because some other studies have not shown such a strong relationship between land use and measures such as soil productivity. There was a degree of natural coincidence between more undulating and forested landscapes and lower productive potential. Notably, landscape attributes appeared to be of secondary importance in determining land use. The presence of a house at the time of sale was not a statistically significant explanatory factor for land use. This finding, together with that of limited conversion of suitable, productive land to unplanned rural living, meant there was a lack of evidence to support a change to the right to have a house on rural allotments. The findings did indicate that the allotment area required to avoid conversion to unplanned rural living was different to the area required for agricultural viability. In the circumstances of this study area and period, 20 ha would have been an acceptable minimum to retain suitable cropping land in productive use. This is compared to the 60 ha suggested as necessary for farm viability. However, the potential for changed economic conditions to reduce the productive value of agricultural land, and increase demand for rural living, makes it appropriate to be cautious about permitted allotment sizes. A detailed conceptual framework informed the selection of observation units and variables for the third stage of analyses. The framework helped to confirm allotment sales as the observation units, and provided the rationale for excluding the difficult to measure influences of a landowner’s attachment to the land, and inertia, as explanatory variables. The high explanatory power of the logit models provides support for the approach, which may be useful in future studies.

In much of Australia and the United States, the almost universal right to have a house on rural allotments has resulted in rural living settlement in areas intended, by land use planning policy, to be used for agriculture. Such ‘unplanned’ rural living has a range of potential adverse impacts, including, among other things, loss of agricultural production, land use conflicts with agriculture, land value constraints on farm restructuring, and changes to the rural landscape character. Given the strongly held community perception of a house as a development right, this research was motivated by the need to understand the effects of that right, in order to assess the potential value of any change in policy. An integral consideration was the influence of the size of allotments with any such right. The research was also driven by an interest in explaining the pattern of unplanned rural living settlement, i.e. why are particular allotments converted and others not? Largely in the absence of previous studies of unplanned rural living as a distinct phenomenon, there was scope for this thesis to make a contribution to knowledge in these respects. The research involved three stages of empirical analysis. Firstly, a Queensland-wide survey of land use conversion from sugarcane production supported the choice of the Bundaberg region as the study area. Contingency table analyses were then conducted into the current land use and other attributes of land in the study area that, in 1980, was used for sugarcane. Those analyses identified distinguishing attributes of the different land use sub-sets of allotments, and informed the selection of observation units for the third stage. The third stage provided the principal basis for fulfilling the research objectives. It focused on logit models of the choice, at the time of sale, between using a particular allotment for unplanned rural living or sugarcane production. Those analyses provided an indication of the role of individual land attributes while controlling for the effects of other attributes. It was found there had been limited conversion of suitable, productive sugarcane land to unplanned rural living. Compared to those allotments that remained in sugarcane production, allotments converted to rural living were, on average, much smaller, of less value, had lower agricultural production potential, and were situated in more undulating and forested landscapes further from Bundaberg. These clear distinctions between rural living and sugarcane allotments contributed to logit models with high explanatory power. The expected productive income of allotments had overwhelming weight and was highly statistically significant in explaining the land use choice between unplanned rural living and sugarcane. This is a notable finding, because some other studies have not shown such a strong relationship between land use and measures such as soil productivity. There was a degree of natural coincidence between more undulating and forested landscapes and lower productive potential. Notably, landscape attributes appeared to be of secondary importance in determining land use. The presence of a house at the time of sale was not a statistically significant explanatory factor for land use. This finding, together with that of limited conversion of suitable, productive land to unplanned rural living, meant there was a lack of evidence to support a change to the right to have a house on rural allotments. The findings did indicate that the allotment area required to avoid conversion to unplanned rural living was different to the area required for agricultural viability. In the circumstances of this study area and period, 20 ha would have been an acceptable minimum to retain suitable cropping land in productive use. This is compared to the 60 ha suggested as necessary for farm viability. However, the potential for changed economic conditions to reduce the productive value of agricultural land, and increase demand for rural living, makes it appropriate to be cautious about permitted allotment sizes. A detailed conceptual framework informed the selection of observation units and variables for the third stage of analyses. The framework helped to confirm allotment sales as the observation units, and provided the rationale for excluding the difficult to measure influences of a landowner’s attachment to the land, and inertia, as explanatory variables. The high explanatory power of the logit models provides support for the approach, which may be useful in future studies.

In much of Australia and the United States, the almost universal right to have a house on rural allotments has resulted in rural living settlement in areas intended, by land use planning policy, to be used for agriculture. Such ‘unplanned’ rural living has a range of potential adverse impacts, including, among other things, loss of agricultural production, land use conflicts with agriculture, land value constraints on farm restructuring, and changes to the rural landscape character. Given the strongly held community perception of a house as a development right, this research was motivated by the need to understand the effects of that right, in order to assess the potential value of any change in policy. An integral consideration was the influence of the size of allotments with any such right. The research was also driven by an interest in explaining the pattern of unplanned rural living settlement, i.e. why are particular allotments converted and others not? Largely in the absence of previous studies of unplanned rural living as a distinct phenomenon, there was scope for this thesis to make a contribution to knowledge in these respects. The research involved three stages of empirical analysis. Firstly, a Queensland-wide survey of land use conversion from sugarcane production supported the choice of the Bundaberg region as the study area. Contingency table analyses were then conducted into the current land use and other attributes of land in the study area that, in 1980, was used for sugarcane. Those analyses identified distinguishing attributes of the different land use sub-sets of allotments, and informed the selection of observation units for the third stage. The third stage provided the principal basis for fulfilling the research objectives. It focused on logit models of the choice, at the time of sale, between using a particular allotment for unplanned rural living or sugarcane production. Those analyses provided an indication of the role of individual land attributes while controlling for the effects of other attributes. It was found there had been limited conversion of suitable, productive sugarcane land to unplanned rural living. Compared to those allotments that remained in sugarcane production, allotments converted to rural living were, on average, much smaller, of less value, had lower agricultural production potential, and were situated in more undulating and forested landscapes further from Bundaberg. These clear distinctions between rural living and sugarcane allotments contributed to logit models with high explanatory power. The expected productive income of allotments had overwhelming weight and was highly statistically significant in explaining the land use choice between unplanned rural living and sugarcane. This is a notable finding, because some other studies have not shown such a strong relationship between land use and measures such as soil productivity. There was a degree of natural coincidence between more undulating and forested landscapes and lower productive potential. Notably, landscape attributes appeared to be of secondary importance in determining land use. The presence of a house at the time of sale was not a statistically significant explanatory factor for land use. This finding, together with that of limited conversion of suitable, productive land to unplanned rural living, meant there was a lack of evidence to support a change to the right to have a house on rural allotments. The findings did indicate that the allotment area required to avoid conversion to unplanned rural living was different to the area required for agricultural viability. In the circumstances of this study area and period, 20 ha would have been an acceptable minimum to retain suitable cropping land in productive use. This is compared to the 60 ha suggested as necessary for farm viability. However, the potential for changed economic conditions to reduce the productive value of agricultural land, and increase demand for rural living, makes it appropriate to be cautious about permitted allotment sizes. A detailed conceptual framework informed the selection of observation units and variables for the third stage of analyses. The framework helped to confirm allotment sales as the observation units, and provided the rationale for excluding the difficult to measure influences of a landowner’s attachment to the land, and inertia, as explanatory variables. The high explanatory power of the logit models provides support for the approach, which may be useful in future studies.

Environmental management in modern industry entails much more than simply measuring the level of contaminants at the discharge point of a chimney stack or pipeline and operating within defined licence limits. It involves: 1) Understanding the environmental risks associated with the industrial operation; 2) Quantifying the environmental assets at risk; 3) Evaluating the environmental challenge from the industry (determining the likely effect of particular emission levels on different environments); 4) Monitoring the condition of environmental assets in response to this challenge; 5) Devising appropriate remedial action where necessary. This study aimed to provide a rigorous and effective framework for decision making concerning vegetated landscapes surrounding industrial premises, particularly those associated with emissions of pollutants to air. Three Alcoa of Australia managed facilities in Victoria were used to develop and test the procedures, namely a coastal site at Point Henry, Geelong, a hinterland forest and heath site at Anglesea and a coastal heath site at Portland. The three facilities were involved in the aluminium smelting industry and the major atmospheric emissions were gaseous and particulate fluorides and sulfur dioxide. Analyses of vegetation distribution and condition were undertaken in order to establish whether the industrial activities at the three sites could be identified as the causes of changes in vegetation. A geographic information system (GIS) was implemented at each facility, to contain cadastral information as well as records of the physical environment and plant and animal species occurrences and condition, where appropriate. The GIS was used to create a surface of vegetation condition over the area of interest at the time of assessment and then over time to evaluate vegetation change and relationship to meteorological and production data. Baseline vegetation condition was established for each facility using large-scale high quality aerial photography and multi-spectral imagery. The photography for each site contained large amounts of latent information on vegetation distribution and condition. When properly rectified and geo-referenced, the images became accessible and open to manipulation within the GIS. At each of the three sites investigated, image classes were selected that provided appropriate detail for the vegetation type. In effect the image became a surrogate of the vegetation frozen in time and space. The ability to extract information on past vegetation condition was shown to be a very valuable asset providing the opportunity to generate new contemporaneous data to augment poor or lost historical data. The process was illustrated by developing vegetation change maps and trend information using past and current photography and limited historical field data. A verified emission model was applied to two of the facilities (Anglesea and Portland) to predict emission effects over the vegetated areas. These were used to evaluate current vegetation condition and change in terms of industrial challenge and other change agencies known or suspected. In each case, further developments of the model were required in order to achieve acceptable predictions of known atmospheric and air pollutant conditions. The investigations at Anglesea demonstrated the importance of fine-scale topographic description and meteorological modelling in order to reconcile modelled and observed environmental conditions in hilly coastal terrain. At Portland, the study revealed the importance of marine aerosols in the distribution and deposition of fluoride in the vicinity of the source. Within the broad framework of vegetation type and general condition at Anglesea and Portland, differences at the micro-scale of individual plants and clumps of plants were examined using multi-spectral imagery. Here differences in image characteristics that related to increased pigmentation and loss of chlorophyll in leaves and increased amounts of dead tissue on plants were evaluated. The change agents for the detected differences within sets and between sets of imagery were analysed and again shown to be mainly non-industrial. Only at Portland did atmospheric emissions from the aluminium smelter play a role at some locations, and then only because the effect was being reinforced by other factors (lack of burning and drought). At Portland, population data for associated fauna and an orchid were examined in relation to the baseline vegetation condition, vegetation change and industrial impact. The distribution and welfare of these species were shown to be generally dependent on vegetation condition and other non-industrial factors. If the vegetation overall was maintained in a condition of vigorous growth and structural diversity, it could be expected that specific dependant flora and fauna would thrive within it. The exception was the orchid, which was shown to be sensitive to emissions and could only thrive in its natural habitat in the vicinity of the smelter if protected by screening vegetation or nurtured by an intensive set of cultural practices that were developed as a result of this study.

Environmental management in modern industry entails much more than simply measuring the level of contaminants at the discharge point of a chimney stack or pipeline and operating within defined licence limits. It involves: 1) Understanding the environmental risks associated with the industrial operation; 2) Quantifying the environmental assets at risk; 3) Evaluating the environmental challenge from the industry (determining the likely effect of particular emission levels on different environments); 4) Monitoring the condition of environmental assets in response to this challenge; 5) Devising appropriate remedial action where necessary. This study aimed to provide a rigorous and effective framework for decision making concerning vegetated landscapes surrounding industrial premises, particularly those associated with emissions of pollutants to air. Three Alcoa of Australia managed facilities in Victoria were used to develop and test the procedures, namely a coastal site at Point Henry, Geelong, a hinterland forest and heath site at Anglesea and a coastal heath site at Portland. The three facilities were involved in the aluminium smelting industry and the major atmospheric emissions were gaseous and particulate fluorides and sulfur dioxide. Analyses of vegetation distribution and condition were undertaken in order to establish whether the industrial activities at the three sites could be identified as the causes of changes in vegetation. A geographic information system (GIS) was implemented at each facility, to contain cadastral information as well as records of the physical environment and plant and animal species occurrences and condition, where appropriate. The GIS was used to create a surface of vegetation condition over the area of interest at the time of assessment and then over time to evaluate vegetation change and relationship to meteorological and production data. Baseline vegetation condition was established for each facility using large-scale high quality aerial photography and multi-spectral imagery. The photography for each site contained large amounts of latent information on vegetation distribution and condition. When properly rectified and geo-referenced, the images became accessible and open to manipulation within the GIS. At each of the three sites investigated, image classes were selected that provided appropriate detail for the vegetation type. In effect the image became a surrogate of the vegetation frozen in time and space. The ability to extract information on past vegetation condition was shown to be a very valuable asset providing the opportunity to generate new contemporaneous data to augment poor or lost historical data. The process was illustrated by developing vegetation change maps and trend information using past and current photography and limited historical field data. A verified emission model was applied to two of the facilities (Anglesea and Portland) to predict emission effects over the vegetated areas. These were used to evaluate current vegetation condition and change in terms of industrial challenge and other change agencies known or suspected. In each case, further developments of the model were required in order to achieve acceptable predictions of known atmospheric and air pollutant conditions. The investigations at Anglesea demonstrated the importance of fine-scale topographic description and meteorological modelling in order to reconcile modelled and observed environmental conditions in hilly coastal terrain. At Portland, the study revealed the importance of marine aerosols in the distribution and deposition of fluoride in the vicinity of the source. Within the broad framework of vegetation type and general condition at Anglesea and Portland, differences at the micro-scale of individual plants and clumps of plants were examined using multi-spectral imagery. Here differences in image characteristics that related to increased pigmentation and loss of chlorophyll in leaves and increased amounts of dead tissue on plants were evaluated. The change agents for the detected differences within sets and between sets of imagery were analysed and again shown to be mainly non-industrial. Only at Portland did atmospheric emissions from the aluminium smelter play a role at some locations, and then only because the effect was being reinforced by other factors (lack of burning and drought). At Portland, population data for associated fauna and an orchid were examined in relation to the baseline vegetation condition, vegetation change and industrial impact. The distribution and welfare of these species were shown to be generally dependent on vegetation condition and other non-industrial factors. If the vegetation overall was maintained in a condition of vigorous growth and structural diversity, it could be expected that specific dependant flora and fauna would thrive within it. The exception was the orchid, which was shown to be sensitive to emissions and could only thrive in its natural habitat in the vicinity of the smelter if protected by screening vegetation or nurtured by an intensive set of cultural practices that were developed as a result of this study.

Environmental management in modern industry entails much more than simply measuring the level of contaminants at the discharge point of a chimney stack or pipeline and operating within defined licence limits. It involves: 1) Understanding the environmental risks associated with the industrial operation; 2) Quantifying the environmental assets at risk; 3) Evaluating the environmental challenge from the industry (determining the likely effect of particular emission levels on different environments); 4) Monitoring the condition of environmental assets in response to this challenge; 5) Devising appropriate remedial action where necessary. This study aimed to provide a rigorous and effective framework for decision making concerning vegetated landscapes surrounding industrial premises, particularly those associated with emissions of pollutants to air. Three Alcoa of Australia managed facilities in Victoria were used to develop and test the procedures, namely a coastal site at Point Henry, Geelong, a hinterland forest and heath site at Anglesea and a coastal heath site at Portland. The three facilities were involved in the aluminium smelting industry and the major atmospheric emissions were gaseous and particulate fluorides and sulfur dioxide. Analyses of vegetation distribution and condition were undertaken in order to establish whether the industrial activities at the three sites could be identified as the causes of changes in vegetation. A geographic information system (GIS) was implemented at each facility, to contain cadastral information as well as records of the physical environment and plant and animal species occurrences and condition, where appropriate. The GIS was used to create a surface of vegetation condition over the area of interest at the time of assessment and then over time to evaluate vegetation change and relationship to meteorological and production data. Baseline vegetation condition was established for each facility using large-scale high quality aerial photography and multi-spectral imagery. The photography for each site contained large amounts of latent information on vegetation distribution and condition. When properly rectified and geo-referenced, the images became accessible and open to manipulation within the GIS. At each of the three sites investigated, image classes were selected that provided appropriate detail for the vegetation type. In effect the image became a surrogate of the vegetation frozen in time and space. The ability to extract information on past vegetation condition was shown to be a very valuable asset providing the opportunity to generate new contemporaneous data to augment poor or lost historical data. The process was illustrated by developing vegetation change maps and trend information using past and current photography and limited historical field data. A verified emission model was applied to two of the facilities (Anglesea and Portland) to predict emission effects over the vegetated areas. These were used to evaluate current vegetation condition and change in terms of industrial challenge and other change agencies known or suspected. In each case, further developments of the model were required in order to achieve acceptable predictions of known atmospheric and air pollutant conditions. The investigations at Anglesea demonstrated the importance of fine-scale topographic description and meteorological modelling in order to reconcile modelled and observed environmental conditions in hilly coastal terrain. At Portland, the study revealed the importance of marine aerosols in the distribution and deposition of fluoride in the vicinity of the source. Within the broad framework of vegetation type and general condition at Anglesea and Portland, differences at the micro-scale of individual plants and clumps of plants were examined using multi-spectral imagery. Here differences in image characteristics that related to increased pigmentation and loss of chlorophyll in leaves and increased amounts of dead tissue on plants were evaluated. The change agents for the detected differences within sets and between sets of imagery were analysed and again shown to be mainly non-industrial. Only at Portland did atmospheric emissions from the aluminium smelter play a role at some locations, and then only because the effect was being reinforced by other factors (lack of burning and drought). At Portland, population data for associated fauna and an orchid were examined in relation to the baseline vegetation condition, vegetation change and industrial impact. The distribution and welfare of these species were shown to be generally dependent on vegetation condition and other non-industrial factors. If the vegetation overall was maintained in a condition of vigorous growth and structural diversity, it could be expected that specific dependant flora and fauna would thrive within it. The exception was the orchid, which was shown to be sensitive to emissions and could only thrive in its natural habitat in the vicinity of the smelter if protected by screening vegetation or nurtured by an intensive set of cultural practices that were developed as a result of this study.

Environmental management in modern industry entails much more than simply measuring the level of contaminants at the discharge point of a chimney stack or pipeline and operating within defined licence limits. It involves: 1) Understanding the environmental risks associated with the industrial operation; 2) Quantifying the environmental assets at risk; 3) Evaluating the environmental challenge from the industry (determining the likely effect of particular emission levels on different environments); 4) Monitoring the condition of environmental assets in response to this challenge; 5) Devising appropriate remedial action where necessary. This study aimed to provide a rigorous and effective framework for decision making concerning vegetated landscapes surrounding industrial premises, particularly those associated with emissions of pollutants to air. Three Alcoa of Australia managed facilities in Victoria were used to develop and test the procedures, namely a coastal site at Point Henry, Geelong, a hinterland forest and heath site at Anglesea and a coastal heath site at Portland. The three facilities were involved in the aluminium smelting industry and the major atmospheric emissions were gaseous and particulate fluorides and sulfur dioxide. Analyses of vegetation distribution and condition were undertaken in order to establish whether the industrial activities at the three sites could be identified as the causes of changes in vegetation. A geographic information system (GIS) was implemented at each facility, to contain cadastral information as well as records of the physical environment and plant and animal species occurrences and condition, where appropriate. The GIS was used to create a surface of vegetation condition over the area of interest at the time of assessment and then over time to evaluate vegetation change and relationship to meteorological and production data. Baseline vegetation condition was established for each facility using large-scale high quality aerial photography and multi-spectral imagery. The photography for each site contained large amounts of latent information on vegetation distribution and condition. When properly rectified and geo-referenced, the images became accessible and open to manipulation within the GIS. At each of the three sites investigated, image classes were selected that provided appropriate detail for the vegetation type. In effect the image became a surrogate of the vegetation frozen in time and space. The ability to extract information on past vegetation condition was shown to be a very valuable asset providing the opportunity to generate new contemporaneous data to augment poor or lost historical data. The process was illustrated by developing vegetation change maps and trend information using past and current photography and limited historical field data. A verified emission model was applied to two of the facilities (Anglesea and Portland) to predict emission effects over the vegetated areas. These were used to evaluate current vegetation condition and change in terms of industrial challenge and other change agencies known or suspected. In each case, further developments of the model were required in order to achieve acceptable predictions of known atmospheric and air pollutant conditions. The investigations at Anglesea demonstrated the importance of fine-scale topographic description and meteorological modelling in order to reconcile modelled and observed environmental conditions in hilly coastal terrain. At Portland, the study revealed the importance of marine aerosols in the distribution and deposition of fluoride in the vicinity of the source. Within the broad framework of vegetation type and general condition at Anglesea and Portland, differences at the micro-scale of individual plants and clumps of plants were examined using multi-spectral imagery. Here differences in image characteristics that related to increased pigmentation and loss of chlorophyll in leaves and increased amounts of dead tissue on plants were evaluated. The change agents for the detected differences within sets and between sets of imagery were analysed and again shown to be mainly non-industrial. Only at Portland did atmospheric emissions from the aluminium smelter play a role at some locations, and then only because the effect was being reinforced by other factors (lack of burning and drought). At Portland, population data for associated fauna and an orchid were examined in relation to the baseline vegetation condition, vegetation change and industrial impact. The distribution and welfare of these species were shown to be generally dependent on vegetation condition and other non-industrial factors. If the vegetation overall was maintained in a condition of vigorous growth and structural diversity, it could be expected that specific dependant flora and fauna would thrive within it. The exception was the orchid, which was shown to be sensitive to emissions and could only thrive in its natural habitat in the vicinity of the smelter if protected by screening vegetation or nurtured by an intensive set of cultural practices that were developed as a result of this study.