Digital soil mapping using reference area and artificial neural networks.

Digital soil mapping is an alternative for the recognition of soil classes in areas where pedological surveys are not available. The main aim of this study was to obtain a digital soil map using artificial neural networks (ANN) and environmental variables that express soillandscape relationships. This study was carried out in an area of 11,072 ha located in the Barra Bonita municipality, state of São Paulo, Brazil. A soil survey was obtained from a reference area of approximately 500 ha located in the center of the area studied. With the mapping units identified together with the environmental variables elevation, slope, slope plan, slope profile, convergence index, geology and geomorphic surfaces, a supervised classification by ANN was implemented. The neural network simulator used was the Java NNS with the learning algorithm "back propagation." Reference points were collected for evaluating the performance of the digital map produced. The occurrence of soils in the landscape obtained in the reference area was observed in the following digital classification: medium-textured soils at the highest positions of the landscape, originating from sandstone, and clayey loam soils in the end thirds of the hillsides due to the greater presence of basalt. The variables elevation and slope were the most important factors for discriminating soil class through the ANN. An accuracy level of 82% between the reference points and the digital classification was observed. The methodology proposed allowed for a preliminary soil classification of an area not previously mapped using mapping units obtained in a reference area

Mapping spatial and temporal potassium balances in Brazilian soils of south-west Goias.

Introduction: Brazil, is one of the main agricultural producers in the world ranking 1st in the production of sugarcane, coffee and oranges. It is also 2nd as world producer of soybeans and a leader in the harvested yields of many other crops. The annual consumption of mineral fertilizers exceeds 20 million mt, 30% of which corresponds to potash fertilizers (ANDA, 2006). From this statistic it may be supposed that fertilizer application in Brazil is rather high, compared with many other countries. However, even if it is assumed that only one fourth of this enormous 8.5 million km2 territory is used for agriculture, average levels of fertilizer application per hectare of arable land are not high enough for sustainable production. One of the major constraints is the relatively low natural fertility status of the soils which contain excessive Fe and Al oxides. Agriculture is also often practised on sandy soils so that the heavy rainfall causes large losses of nutrients through leaching. In general, nutrient removal by crops such as sugarcane and tropical fruits is much more than the average nutrient application via fertilization, especially in regions with a long history of agricultural production. In the recently developed areas, especially in the Cerrado (Brazilian savanna) where agriculture has expanded since 1980, soils are even poorer than in the "old" agricultural regions, and high costs of mineral fertilizers have become a significant input factor in determining soybean, maize and cotton planting. The consumption of mineral fertilizers throughout Brazil is very uneven. According to the 1995/96 Agricultural Census, only in eight of the total of 26 Brazilian states, were 50 per cent or more of the farms treated "systematically" with mineral fertilizers; in many states it was less than 25 per cent, and in five states even less than 12 per cent (Brazilian Institute for Geography and Statistics; Censo Agropecuario1995/96, Instituto Brazileiro de Geografia e Estadistica; IBGE, www.ibge.gov.br). The geographical application distribution pattern of mineral fertilizers may be considered as an important field of research. Understanding geographical disparities in fertilization level requires a complex approach. This includes evaluation of the availability of nutrients in the soil (and related soil properties e.g. CEC and texture), the input of nutrients with fertilizer application, and the removal of nutrients by harvested yields. When all these data are compiled, it is possible to evaluate the balance of particular nutrients for certain areas, and make conclusions as to where agricultural practices should be optimized. This kind of research is somewhat complicated, because it relies on completely different sources of data, usually from incomparable data sources, e.g. soil characteristics attributed to soil type areas, in contrast to yields by administrative regions, or farms. A priority tool in this case is the Geographical Information System (GIS), which enables attribution of data from different fields to the same territorial units, and makes possible integration of these data in an "inputoutput" model, where "input" is the natural availability of a nutrient in the soil plus fertilization, and "output" export of the same nutrient with the removed harvested yield.