There is a recognized importance of soil organic carbon (SOC) and its profound influence on soil properties and crucial part of the global carbon cycle. Organic carbon is a single component of soil constituents and its mapping should take into account its compositional nature. The statistical analysis of compositional data sets is based on the Aitchison's idea (1986) and its developments that if all chemical elements in a soil sample are analysed, their concentrations sum up to a constant and the relationships that variables have to one another are conditioned from the rest of the elements included in the composition. The study was carried out within the project AlForlab PON03PE_00024_1, and aimed to map soil organic carbon within a forested catchment in southern Italy accounting for its compositional nature. The study area is a 139 ha catchment on granitic parent material and subordinately alluvial deposits, where soils are classified as Typic Xerumbrepts and Ultic Haploxeralf crop out. Soils samples were collected at 135 locations and the sample design was developed using a spatial simulated annealing algorithm with two steps. First, one half of the samples were located over the whole basin onto a 5 m grid, optimizing for minimal distance between observations. Then, the remaining samples were located using the Weighted Means of Shortest Distances algorithm, with slope gradient as a weigh function. In the field, soils were sampled up to a depth of 0.20 m and the geographical coordinates of each point were recorded with a differential Trimble receiver at 1-m accuracy. In the laboratory, SOC concentration was measured using a Shimadzu TOC-L analyzer with a SSM-5000A solid sample module whereas through sieve and hydrometer methods, the percentage of sand, silt and clay was determined. A geostatistical approach was used to map SOC concentration at the 0.20 m depth over the whole catchment. A complementary variable (filler) was calculated to complete the composition and to ensure constant sum of SOC interpolated values at a given location. The isometric logratio (irl) transformation of raw data was used. Moreover, to assess the contribution of soil textural fractions in improving the estimation of SOC concentration at unsampled locations, sand, silt and clay concentrations were used as auxiliary information within ordinary cokriging. The three textural components were used as both raw data and isometric logratio (irl) transformations. Four approaches were compared to map SOC using (1) only raw SOC data concentrations, (2) isometric logratio transformation of SOC and the complementary variable, (3) raw data both for SOC and textural components, and (4) isometric logratio transformation of SOC and textural components. The comparison was evaluated splitting the SOC data randomly into calibration and validation sets. Since textural data were used as auxiliary variables in the estimation procedure, whole data set was used. In both cases (univariate and multivariate), using isometric logratio transformations compared to raw data was more computationally intensive and in estimating SOC concentrations, slightly better results were obtained.
ACCOUNTING FOR COMPOSITIONAL NATURE OF SOIL COSTITUENTS IN DIGITAL MAPPING OF ORGANIC CARBON
Gabriele Buttafuoco;Donatella Civitelli;Massimo Conforti;Anna Lia Gabriele
2021
Abstract
There is a recognized importance of soil organic carbon (SOC) and its profound influence on soil properties and crucial part of the global carbon cycle. Organic carbon is a single component of soil constituents and its mapping should take into account its compositional nature. The statistical analysis of compositional data sets is based on the Aitchison's idea (1986) and its developments that if all chemical elements in a soil sample are analysed, their concentrations sum up to a constant and the relationships that variables have to one another are conditioned from the rest of the elements included in the composition. The study was carried out within the project AlForlab PON03PE_00024_1, and aimed to map soil organic carbon within a forested catchment in southern Italy accounting for its compositional nature. The study area is a 139 ha catchment on granitic parent material and subordinately alluvial deposits, where soils are classified as Typic Xerumbrepts and Ultic Haploxeralf crop out. Soils samples were collected at 135 locations and the sample design was developed using a spatial simulated annealing algorithm with two steps. First, one half of the samples were located over the whole basin onto a 5 m grid, optimizing for minimal distance between observations. Then, the remaining samples were located using the Weighted Means of Shortest Distances algorithm, with slope gradient as a weigh function. In the field, soils were sampled up to a depth of 0.20 m and the geographical coordinates of each point were recorded with a differential Trimble receiver at 1-m accuracy. In the laboratory, SOC concentration was measured using a Shimadzu TOC-L analyzer with a SSM-5000A solid sample module whereas through sieve and hydrometer methods, the percentage of sand, silt and clay was determined. A geostatistical approach was used to map SOC concentration at the 0.20 m depth over the whole catchment. A complementary variable (filler) was calculated to complete the composition and to ensure constant sum of SOC interpolated values at a given location. The isometric logratio (irl) transformation of raw data was used. Moreover, to assess the contribution of soil textural fractions in improving the estimation of SOC concentration at unsampled locations, sand, silt and clay concentrations were used as auxiliary information within ordinary cokriging. The three textural components were used as both raw data and isometric logratio (irl) transformations. Four approaches were compared to map SOC using (1) only raw SOC data concentrations, (2) isometric logratio transformation of SOC and the complementary variable, (3) raw data both for SOC and textural components, and (4) isometric logratio transformation of SOC and textural components. The comparison was evaluated splitting the SOC data randomly into calibration and validation sets. Since textural data were used as auxiliary variables in the estimation procedure, whole data set was used. In both cases (univariate and multivariate), using isometric logratio transformations compared to raw data was more computationally intensive and in estimating SOC concentrations, slightly better results were obtained.File | Dimensione | Formato | |
---|---|---|---|
prod_468998-doc_189747.pdf
solo utenti autorizzati
Descrizione: ACCOUNTING FOR COMPOSITIONAL NATURE OF SOIL COSTITUENTS IN DIGITAL MAPPING OF ORGANIC CARBON
Tipologia:
Versione Editoriale (PDF)
Licenza:
NON PUBBLICO - Accesso privato/ristretto
Dimensione
89.58 kB
Formato
Adobe PDF
|
89.58 kB | Adobe PDF | Visualizza/Apri Richiedi una copia |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.