Garth Warren

Reversing a tree regeneration crisis in an endangered ecoregion

Global food demand is growing rapidly. Livestock grazing can provide a valuable source of protein, but conventional grazing is often unsustainable. We studied an 800,000-ha section of a threatened ecoregion in southeastern Australia. Conventional management in the region involves continuous livestock grazing with few rest periods and regular fertilizer application. By using remotely sensed data on tree cover and extensive field data on livestock grazing regimes, soil chemistry, tree diameters, and tree regeneration, we show that the region is facing a tree regeneration crisis. Under conventional management, across the region, millions of hectares of land currently supporting tens of millions of trees will be treeless within decades from now. This would have severe negative ramifications for biodiversity and key ecosystem services, including water infiltration and shade provision for livestock. However, we identified an unexpected win–win solution for tree regeneration and commercial grazing. A relatively new practice in the region is fast-rotational grazing, characterized by prolonged rest periods in between short, intensive grazing events. The probability of regeneration under fast-rotational grazing was up to 4-fold higher than under conventional grazing, and it did not differ significantly from the probability of regeneration in ungrazed areas. In addition, trees were more likely to regenerate where soil nutrient levels were low. These findings suggest that the tree regeneration crisis can be reversed by applying low-input, fast-rotational grazing. New policy settings supporting these practices could signal a turning point for the region, from ecological decline to ecological recovery.

Toward landscape-wide conservation outcomes in Australia's temperate grazing region.

Agriculture and livestock grazing threaten biodiversity around the world. In the grazing landscapes of eastern Australia, a common conservation strategy has been to exclude livestock from large patches of trees (typically > 5 ha). This has major local benefits, but is unlikely to stem regional biodiversity loss. Using a case study from the Upper Lachlan catchment in New South Wales, we show that (1) approximately 30% of tree cover occurs as very small patches or scattered trees; (2) large patches have disappeared from 90% of the landscape; and (3) large patches are 3.5 times more likely to be in unproductive upland areas than in lowland areas of high conservation concern. Given the limitations of focusing on large patches of trees to achieve regional conservation outcomes, the next generation of conservation initiatives should consider a new suite of additional measures that could deliver biodiversity benefits across broad areas of the region. Two key measures that must be considered are new incentives fo...

Multi-criteria assessment for linking regional conservation planning and farm-scale actions

Regional-scale ecological restoration priorities such as increasing the extent and quality of native vegetation are generally planned at catchment scales, while on-ground restoration actions are generally implemented at paddock or farm scales. This paper describes the use of spatial multi-criteria assessment methodologies to construct maps of regional conservation priorities and assesses how these maps map influence farm-scale actions in Western Victoria, Australia (e.g. farm-scale revegetation for salinity, wind erosion, stock shelter, etc). The study also incorporates agricultural production in the decision analysis through the use of historical yield mapping data obtained from harvest logs from precision agriculture equipment. Via a stakeholder workshop, farmer land use priorities were elicited with and without access to maps of regional conservation priorities. Results highlight that production imperatives drive farmer-led conservation actions and that regional conservation priorities have only limited impact on actions. The paper also identifies limitations of applying MCA methods across multiple decision-making scales such as the need to generalise priorities where domain knowledge is relatively high, and the challenges associated with MCA criteria definition.

Large-scale regional delineation of riparian vegetation in the arid and semi-arid Pilbara region, WA

Multi-scene Landsat 5 TM imagery, Principal Component Analysis (PCA), and the Normalised Difference Vegetation Index (NDVI) were used to produce the first region-scale map of riparian vegetation for the Pilbara (230,000 km2), Western Australia. Riparian vegetation is an environmentally important habitat in the arid and desert climate of the Pilbara. These habitats are supported by infrequent flow events and in some locations by groundwater discharge. Our analysis suggests that riparian vegetation covers less than 4% of the Pilbara region, while almost 10.5% of this area is comprised of Groundwater Dependent Vegetation (GDV). GDV is often associated with open water (river pools), providing refugia for a variety of species. GDV has an extremely high ecological value and are often important Indigenous sites. This paper demonstrates how Landsat data calibrated to Top of Atmosphere (TOA) reflectance can be used to delineate riparian vegetation across 16 Landsat Scenes and two UTM (Universal Transverse Mercator) spatial zones. The proposed method is able to delineate riparian vegetation and GDV, without the need for Bidirectional Reflectance Distribution Function (BRDF) correction. Results were validated using ground-truth data from local and regional scale vegetation surveys.

Native Vegetation Condition: Site to Regional Assessments

There is an increasing emphasis in Australia on the use of vegetation condition information for regional conservation planning. The use of site-based vegetation condition assessments is a relatively mature management application with methods developed for different landscapes (e.g. rangelands and riparian) and to support a variety of natural resource management requirements (property vegetation planning or market-based instruments). On the other hand, the creation of regional-scale maps of native vegetation condition is still a developing methodology. This chapter argues that regional-scale maps of native vegetation condition are an important tool to complement site-based assessments. When combined they can provide a powerful integrated tool for regional conservation planning. Through a case study we describe a methodology for extending site-based data to maps of two vegetation condition attributes based on the BioMetric site assessment method. The case study, in the Murray Catchment in New South Wales, Australia, illustrates how an understanding of faunal response to native vegetation condition can be combined with modelled data to develop regional conservation planning maps. A spatial data aggregation approach is applied to model outcomes to address concerns about uncertainty and data confidentiality.