Avaliaram-se séries climáticas (1960-91) de precipitação e temperatura do ar mensal de produtos em grade em relação às séries desses elementos observadas em estações meteorológicas do estado do Rio de Janeiro. As séries climáticas observadas foram obtidas nas estações do Instituto Nacional de Meteorologia, localizadas no estado do Rio de Janeiro. As séries dos produtos em grade foram extraídas nos pontos de grade (resolução 0,5 o x 0,5o) dos produtos do Global Precipitation Climatology Center (GPCC), The Global Historical Climatology Network (GHCN) ou Universidade de Delaware (UDEL) mais próximos das estações em estudo. Avaliou-se a precisão (coeficiente de determinação – r²) e exatidão (índice de concordância de Willmott – d e Raiz do Quadrado Médio do Erro - RQME) de cada produto em grade em relação às séries observadas. Os produtos em grade de precipitação (GPCC e UDEL) não tiveram precisão satisfatória (r² < 0,54 - GPCC e r² < 0,61 - UDEL), contudo sua exatidão (d > 0,59 - GPCC e d > 0,58 - UDEL) foi superior à precisão. Os erros observados para precipitação foram entre 59,5 e 125,8 mm. As séries em grade de temperatura tiveram maior precisão (r² > 0,41 - GHCN e r² > 0,35 - UDEL) e exatidão similar (d > 0,58 - GHCN e d > 0,65 - UDEL), com RQME entre 1,11 e 3,98 oC. Foram identificadas associações que elevam o erro dos produtos em grade para a região, tais como, o elevado gradiente altitudinal da área de estudo e o efeito continentalidade/maritimidade. Dentre os produtos em grade de precipitação, o GPCC apresentou melhor desempenho (maior precisão e exatidão) em relação à UDEL na maior parte das estações, enquanto para temperatura do ar, as séries em grade da UDEL se sobressaíram em comparação ao GHCN. É necessário desenvolver produtos climáticos de precipitação e temperatura do ar em grade precisos e exatos com alta resolução para o estado do Rio de Janeiro.

No abstract

The aim of this work was to propose a method for the consistency of climatic series of monthly rainfall using a supervised and unsupervised approach. The methodology was applied for the series (1961-2010) of rainfall from weather stations located in the State of Rio de Janeiro (RJ) and in the borders with the states of São Paulo, Minas Gerais and Espírito Santo with the State of Rio de Janeiro. The data were submitted to quality analysis (physical and climatic limit and, space-time tendency) and gap filling, based on simple linear regression analysis, associated with the prediction band (p < 0.05 or 0.01), in addition to the Z-score (3, 4 or 5). Next, homogeneity analysis was applied to the continuous series, using the method of cumulative residuals. The coefficients of determination (r²) between the assessed series and the reference series were greater than 0.70 for gap filling both for the supervised and unsupervised approaches. In the analysis of data homogeneity, supervised and unsupervised approaches were effective in selecting homogeneous series, in which five out of the nine final stations were homogeneous (p > 0.9). In the other series, the homogeneity break points were identified and the simple linear regression method was applied for their homogenization. The proposed method was effective to consist of the rainfall series and allows the use of these data in climate studies.

This chapter discusses urban mobility considering two main analyses approaches. Based on the relationship between mobility and vulnerability, the first approach analyzed commuter's vulnerability using basin as unit of analysis. The second one analyzes variables related to land use such as population density and its relation with job offer in the city and people's income using traffic zones as unit of analysis. The two scales dialogue and can be used concurrently. The municipality of São José dos Campos (Brazil) was used as a case study. Origin-destination research was the main database used in the analyses. Authors used geospatial tools, like spatial join operation and thematic maps, which enable the in-depth analysis of important data for urban studies or transport planning and can be replicated in any study area. The analysis of mobility data aggregated by basin contributed to an understanding of the implications of the urban configuration, with its displacement patterns related to water courses if any flooding or landslide occurs and interrupts people's flow.

Desertification is a land degradation phenomenon with dire and irreversible consequences, affecting different regions of the world. Assessment of spatial susceptibility to desertification requires long-term series of precipitation (P) and evapotranspiration (PET). An approach to desertification analysis is the use of spatially gridded time series of air temperature and precipitation, derived from spatial interpolation of in situ measurements and available globally. The aim of this article was to estimate the susceptibility to desertification over Southeast Brazil using monthly gridded data from the Global Precipitation Climatology Centre (GPCC), and from the Global Historical Climatology Network (GHCN). Two indices were used to estimate desertification susceptibility: the aridity index Ia (P/PET) and D (PET/P). Validation of these datasets was performed using in situ observations (1961—2010) from the National Institute of Meteorology (INMET) – (68 weather stations). Determination coefficient (r²) and the Willmott’s coefficient of agreement (d) between gridded and observed data revealed satisfactory accuracy and precision for grids of precipitation (r2 > 0.93, d > 0.90), air temperature (r2 > 0.94, d > 0.53) and PET (r2 > 0.93, d > 0.63). Areas susceptible to desertification were identified by the index Ia over the Northern regions of Minas Gerais and Rio de Janeiro states. No areas susceptible to desertification were identified using the index D. However, both indices indicated large areas of dry sub-humid climate, which can be strongly affected by drought conditions. Overall, climate gridded variables presented good precision and accuracy when used to identify areas susceptible to desertification.