Mathematical modeling and use of orbital products in the environmental degradation of the Araripe Forest in the Brazilian Northeast

• Published in: Modeling earth systems and environment 2019-12, Vol.5 (4), p.1429-1441
• Main Authors: Barros Santiago, Dimas de, Correia Filho, Washington Luiz Félix, de Oliveira-Júnior, José Francisco, da Silva Junior, Carlos Antonio
• Format: Article
• Language: English
• Subjects: Arid regions; Chemistry and Earth Sciences; Computer Science; Earth and Environmental Science; Earth Sciences; Earth System Sciences; Ecosystems; Environment; Environmental degradation; Forests; Land surface temperature; Landsat; Landsat satellites; Livestock; Math. Appl. in Environmental Science; Mathematical Applications in the Physical Sciences; Mathematical models; Modelling; Moisture index; Normalized difference vegetative index; Original Article; Physics; Plant cover; Products; Rain; Rainfall; Regression models; Remote sensing; Soil; Soil moisture; Statistics for Engineering; Surface temperature; Vegetation; Vegetation cover; Vegetation index
• ISSN: 2363-6203; 2363-6211
• DOI: 10.1007/s40808-019-00614-x

Vegetation cover is indispensable in the process of inhibiting environmental degradation. In the Northeast of Brazil, especially in the Araripe Nacional Forest (FLONA), this problem is related to the removal of vegetation for industrial and domestic use, in addition to the expansion of livestock. Thus, the objective of this work was to evaluate the environmental degradation in the area of FLONA from orbital products via remote sensing with the aid of mathematical modeling. For this, two orbital images of the orbit 65, point 217 were used for processing and obtaining the variables: (1st) July 7, 2003 from TM/Landsat-5 and (2nd) July 15, 2018 from OLI/Landsat-8. In mathematical modeling, the multiple linear regression (MRL) model was applied to the orbital products: land surface temperature (LST), normalized burn ratio (NBR), Normalized Difference Moisture Index, Normalized Difference Water Index (NDWI) to estimate the Soil Adjusted Vegetation Index (SAVI) and hence to predict the Normalized Difference Vegetation Index (NDVI). All the processing to obtain the results was carried out in the software R version 3.4-1. O NDVI pointed out a significant increase of 72.05% in dense vegetation, from 158.33 to 272.40 km 2 . However, vegetation is more likely to suffer from stress due to the increase in LST at 5 °C, which increased from 17.5 to 25.0 °C, reaching its highest value of 42 °C in July 2011. The MRL results indicated that the models have an excellent predictive capacity in the estimation of degradation, with R 2 value greater than 92% of the explained variance. In addition, the MAE and root mean square error were less than 0.03 for both models. The models pointed out that SAVI, NBR and NDWI are responsible for the variability of NDVI in environmental degradation of FLONA. Highlight for the theoretical-conceptual model that can be applied to any semi-arid and highly-sensitive region to changes in the rainfall pattern.