Diagnostic soil attributes to predict and manage soil corrosion damage to exclusion fencing used in conservation programs

The use of exclusion fencing as part of wildlife conservation programs has been increasing in recent years, particularly in Australia. Soil corrosion damage sustained on fences is a significant management concern as the weakened fence netting can provide opportunities for feral animal incursions into fenced safe havens. Soil corrosivity risk mapping can assist with the design of fenced nature reserves to reduce the frequency of fence repair and replacement. However, very little research has focused on developing methods for accurately predicting fence corrosion rates in different surface soil environments. This paper assesses the use of different soil attributes as corrosivity indicators for identifying areas of low, moderate and high fence corrosion risk in different soil environments present in South Australia (20 field sites). Zinc corrosion rates measured on zinc-aluminium fence samples (buried at sites for 9 months) ranged by a factor of nearly 50, with low rates of fence corrosion (0.1-0.7 mu m/year) observed at five sites, medium rates (0.7-2.1 mu m/year) observed at 10 sites, and extreme rates (>8.4 mu m/year) observed at four sites. Fence corrosion risk was predicted using soil pH, soil salinity and texture data, and a soil corrosivity risk index developed for use in arid soils in South Australia. Predicted zinc corrosion rates matched field observations at 45 % of field sites. The highest rates of zinc corrosion (>4.2 mu m/year) were observed at field sites with highly alkaline (pH > 8.5) and highly saline (ECe >= 5 dS/m) soils. An improved fence corrosion risk classification method, referred as the Fence Corrosion Risk Decision Tree was developed using these soil pH and salinity thresholds, which correctly predicted fence corrosion risk at 67 % of field sites at Olympic Dam and Farina and 50 % of field sites on the Yorke Peninsula. Further research is needed to assess the ability of this method to predict long-term fence damage (>2 years exposed to soil conditions).