Urban soils as a spatial indicator of quality for urban socio-ecological systems.

The development of criteria and indicators to quantify the transition to sustainability of the urban socio-ecological systems quality is determinant for planning policies and the 21st century urban agenda. This study models the spatial variation in the concentration and distribution of some macronutrients, micronutrients, and trace nutrients in the soil of a high-altitude city in the Andes. Meanwhile, machine learning methods were employed to study some interactions between the different dimensions that constitute an urban socio-ecosystem that caused these variations. We proposed a methodology that considered two phases: a) field work to collect data on 300 soil samples; laboratory analysis to measure the concentrations of 24 macronutrients, micronutrients, and trace nutrients; and the design of geophysical, spectral, and urban co-variables; b) statistical and geo-informatics analysis, where multivariate analysis grouped the elements into factors; and, machine learning integrated with co-variables was applied to derive the intensity of each factor across the city. Multivariate statistics described the variation in soil co-concentrations with a moderate percentage (42%). Four factors were determined that grouped some of the analyzed elements, as follows: F1 (Zn, S, Cu, Pb, Ni, and Cr), F2 (Ba, Ag, K, In, and Mg), F3 (B, V, Li, and Sr), and F4 (Si and Mn). The percentage R2 out-of-bag of the spatial model were: F1 = 20%, F2 = 8%, F3 = 14%, and F4 = 10%. Our outputs show that the enrichment and contamination by anthropogenic factors, such as the increase in population density, land use, road network, and traffic generated by fossil fuel vehicles, should be prioritized in urban planning decisions.

Space-time analysis of vegetation trends and drought occurrence in domain area of tropical forest.

The purpose of this study is to evaluate temporal trends in changes in vegetation patterns within the Sooretama Biological Reserve and its surroundings, located in Espirito Santo State, Brazil. The evaluation will be performed using the EVI and NDVI index of the MODIS sensor, the Mann-Kendall monotonic trend, Seasonal Trend Analysis methods, and monitoring drought events through the VCI drought index for the years 2007 through 2015. The tools utilized were the EVI and NDVI indexes of the MOD13Q1 product and LST from the MOD11A2 product. These indices were used in order to represent the dynamics of the study area biomass and then to analyze the drought occurrence using the index best-suited to the area of study, identified as VCI. The temporal trends in the data set were examined, pixel by pixel, by application of the Mann-Kendall monotonic technique, treating each pixel in space as a one-dimensional temporal series of 16-day cycles. To evaluate the seasonal trend, the analysis used the STA technique (Seasonal Trend Analysis) implemented in the ETM module. The characterization and spatial distribution of drought events were performed through the Vegetation Condition Index (VCI). The use of (a) images and seasonal curves produced by the monotonic trend of Mann-Kendall and (b) analysis of seasonal trends generated the response of the vegetation to climate variations. The VCI indicated a potential for drought occurrence analysis in regions and areas with different vegetation densities. So, the VCI can be used as a powerful tool to compose a comprehensive and early system alert of drought that can accompany the changes in spatial coverage of vegetation and severity of change. Lastly, the analysis of the data from the MODIS NDVI, EVI, and TST images indicated that the data is suitable to a space-time analysis of drought occurrences and vegetation trends.