Neural network integration of field observations for soil endocrine disruptor characterisation.

A neural network approach was used to predict the presence and concentration of a range of endocrine disrupting compounds (EDCs), based on field observations. Soil sample concentrations of endocrine disrupting compounds (EDCs) and site environmental characteristics, drawn from the National Soil Inventory of Scotland (NSIS) database, were used. Neural network models were trained to predict soil EDC concentrations using field observations for 184 sites. The results showed that presence/absence and concentration of several of the EDCs, mostly no longer in production, could be predicted with some accuracy. We were able to predict concentrations of seven of 31 compounds with r(2) values greater than 0.25 for log-normalised values and of eight with log-normalised predictions converted to a linear scale. Additional statistical analyses were carried out, including Root Mean Square Error (RMSE), Mean Error (ME), Willmott's index of agreement, Percent Bias (PBIAS) and ratio of root mean square to standard deviation (RSR). These analyses allowed us to demonstrate that the neural network models were making meaningful predictions of EDC concentration. We identified the main predictive input parameters in each case, based on a sensitivity analysis of the trained neural network model. We also demonstrated the capacity of the method for predicting the presence and level of EDC concentration in the field, identified further developments required to make this process as rapid and operator-friendly as possible and discussed the potential value of a system for field surveys of soil composition.

Controls on soil solution nitrogen along an altitudinal gradient in the Scottish uplands.

Nitrogen (N) deposition continues to threaten upland ecosystems, contributing to acidification, eutrophication and biodiversity loss. We present results from a monitoring study aimed at investigating the fate of this deposited N within a pristine catchment in the Cairngorm Mountains (Scotland). Six sites were established along an elevation gradient (486-908 m) spanning the key habitats of temperate maritime uplands. Bulk deposition chemistry, soil carbon content, soil solution chemistry, soil temperature and soil moisture content were monitored over a 5 year period. Results were used to assess spatial variability in soil solution N and to investigate the factors and processes driving this variability. Highest soil solution inorganic N concentrations were found in the alpine soils at the top of the hillslope. Soil carbon stock, soil solution dissolved organic carbon (DOC) and factors representing site hydrology were the best predictors of NO(3)(-) concentration, with highest concentrations at low productivity sites with low DOC and freely-draining soils. These factors act as proxies for changing net biological uptake and soil/water contact time, and therefore support the hypothesis that spatial variations in soil solution NO(3)(-) are controlled by habitat N retention capacity. Soil percent carbon was a better predictor of soil solution inorganic N concentration than mass of soil carbon. NH(4)(+) was less affected by soil hydrology than NO(3)(-) and showed the effects of net mineralization inputs, particularly at Racomitrium heath and peaty sites. Soil solution dissolved organic N concentration was strongly related to both DOC and temperature, with a stronger temperature effect at more productive sites. Due to the spatial heterogeneity in N leaching potential, a fine-scale approach to assessing surface water vulnerability to N leaching is recommended over the broad scale, critical loads approach currently in use, particularly for sensitive areas.