Soil carbon sequestration - Evaluation of its potential in pastures on sandy soils, Anhembi - Piracicaba/SP / Seqüestro de carbono nos solos - Avaliação das potencialidades dos solos arenosos sob pastagens, Anhembi - Piracicaba - SP"

AUTOR(ES)
DATA DE PUBLICAÇÃO

2003

RESUMO

The project studied the potential for soil carbon sequestration in pastures growing on sandy soils in the Cerrado region and for the application of remote sensing techniques for estimating carbon sequestration levels. The main focus was deducing how many tons per hectare could be sequestered if better pasture management techniques were adopted. Four pastures with identical pedological, climate and topographical characteristics were chosen, so that the discrepancies in carbon stock levels in the soil could be interpreted in light of their different types of management and productivity levels. The pastures are located in the western part of the Piracicaba basin, in São Paulo State, Brazil. The dominant land use in this region is pasture on sandy soils, classified as Neossolos Quartzarênicos (RQ), according to Brazilian soil classification. After the selection phase, the four pastures were classified according to their soil carbon stocks (0-50cm layer) and their respective productivity levels. Pasture productivity was determined by the following parameters: (i) leaf area index (LAI), (ii) biomass growth rate, (iii) dry matter weight and (iv) root stocks. The correlation between soil carbon stocks and each of the productivity parameters was investigated. The productivity levels were also studied in relation to the pasture management history of each of the selected pastures. The results indicate a good correlation between the productivity level of the pastures and their carbon stocks in the soil. The LAI parameter showed a particularly good correlation with soil carbon stocks (r= 0,97). The second objective of the project was investigating the applicability of remote sensing techniques to map the potential for soil carbon sequestration. For the purposes of this study, the pasture with the highest soil carbon stocks (and higher productivity) was considered to represent the maximum stock possible under the climate, pedological and topography conditions found in the region of the study. The difference between the stock in the highest stock pasture and the three other pastures was calculated. This difference represents the potential for increasing carbon sequestration. The total soil carbon sequestration potential in the most degraded pasture (Descalvado) is equivalent to a 70,0% increase (22.41 Mg ha -1 ) from current levels. In the significantly degraded pasture (Bondade), the potential amounted to 30.08% (12.81 Mg ha -1 ) and to 4.9% (22.41Mg ha -1 ) in the moderately degraded pasture (Barreiro Rico). Since the only determinant parameter for soil carbon sequestration these pasture do not have in common is the type of management, the differences in stock levels are directly related to their type of management. Soil carbon stock levels are a subterraneous type of information, and thus cannot be examined through direct use of optical spectrum bands, which cannot go through the soil. Therefore, and taking into consideration that this study demonstrated a good correlation between productivity and soil carbon stocks in the pastures of the sample, the author attempted to map carbon stocks through pasture productivity. To this end, a radiometric classification of pasture productivity was established. Six spectral bands from the ETM+ satellite sensor and the Normalized Difference Vegetation Index (NDVI) were processed in all possible entry band combinations for two supervised classifications (Maximum Likelihood and Back Propagation). An error matrix was subsequently used to determine the best band combinations for distinguishing pasture productivity levels (two classes). The best result (B2 B5 B7 NDVI) revealed an overall accuracy of 64,79% in the classification of two classes of productivity, which is considered insufficient. Nevertheless, a spectrum radiometric analysis of the pastures was conducted, which showed better separability in the medium infrared spectrum than in the visible or near infrared spectrum. Regression analyses were performed between the reflectance of the four pastures under each spectrum and their respective soil carbon stocks levels, with the purpose of investigating whether there are any direct correlations between the spectral behavior of the pastures and the soil carbon stocks levels. The red (r=0.95) and middle infrared (r=0,944) spectra were demonstrated to have a significant correlation with soil carbon stocks. The author concludes that there is a good correlation between pasture management (productivity) and soil carbon stocks. It has also been determined that the satellite image used in this study (Landsat 7) does not have an adequate spacial and spectral resolution for accurately separating different productivity levels. However, good separability of productivity levels was shown in the middle infrared spectrum. The evaluation of a direct correlation between the spectral behavior of the pastures and their respective soil carbon stock levels obtained significant results in the red (690nm) and near infrared (1550- 1750nm) spectra. Therefore, it is recommended to investigate carbon soil stock mapping by remote sensing techniques directly through the spectral reflectance of pastures in the red and middle infrared spectra or indirectly through the LAI, which can be deduced through the NDVI in satellite images.

ASSUNTO(S)

fixação de carbono sensoriamento remoto greenhouse effect carbon fixation remote sensing efeito estufa degradação do solo soil degradation

ACESSO AO ARTIGO

Documentos Relacionados