Variazioni di carbonio organico nei suoli italiani dal 1979 al 2008

Research output: Contribution to journalArticlepeer-review

Abstract

[automatically translated] The soils contain about three times the amount of carbon available worldwide in the vegetation and about twice that present in the atmosphere. However, the organic soil carbon (SOC) declined in many areas, whereas it was detected an increase in atmospheric CO2. Recent research has shown that the changes were of use and management of soil to cause the largest losses of SOC in the recent past, rather than the higher temperatures and changes in precipitation resulting from climate change. The main purpose of this work is to estimate the changes in the organic carbon content of soils (carbon stock, CS) in Italy during the last three decades (1979-2008) and tie it to changes in land use. The study also has as its goal to study the relationship between SOC and the factors of soil formation (soil and climate, morphology, lithology and land use). The Data Bank Soil Survey of Italy was the main source of information. The CS was calculated from the data of the SOC, apparent density and skeleton, which were referred to the first 50 cm of soil, obtaining a single value for each punctual observation by means of weighted average on the basis of the depth of the horizons. A series of geographical attributes have been used to spatialise the timely, especially the DEM (100 m) and derived morphological classes SOTER, the Soil Region of Italy (Reference 1: 5.000000), the lithological groups of systems Terre Italian (reference scale of 1: 500,000) moisture regimes and soil temperature (raster maps with pixels of 1 km), land use (project Corine land cover, reference scale of 1: 100,000; Corine 2009) at maturity dates in 1990 and 2000 and a card land use updated to 2008 from that in 2000, using ground observation points. The interpolation method used was the multiple linear regression (MLR), with CS as a target variable and geographic attributes as predictors. Basic statistical analysis was conducted to individually investigate the relationships between predictor variables considered and the CS. Finally it found a general model of multiple linear regression, considering together all the predictor variables. The best predictor variables were chosen with a step-wise regression, using the Akaike Information Criterion (AIC) as a criterion for selecting the best variables and the best final model. The final model obtained considered the following predictor variables: i) the decades, ii) land use, iii) the SOTER morphological classes, iv) Soil Region, v) the soil temperature regimes, there) schemes soil moisture, vii) the lithological groups of Terre Systems, viii) the soil temperature, ix) soil aridity index (days of dry soil), former) share. In the model it was also considered the interaction of the decade and the land use. The results indicate that the CS is highly correlated with the major groups of land use (forests, pastures, agricultural areas), with moisture regimes and soil temperature, with the lithology, with the morphological classes,
Original languageItalian
Pages (from-to)45-54
Number of pages10
JournalBIOLOGI ITALIANI
Volume10
Publication statusPublished - 2010

Cite this