The arborization of the streets is a valuable asset to the cities. His incalculable benefits demonstrate the importance of trees to the quality of life and environment in cities. In the planning of arborization, the qualitative and quantitative evaluation is an activity of great importance for the understanding of that heritage. Generally to conduct a tree inventory using conventional numerous variables collected in the field that demands high public expenses, from the shift in the streets and forming a team, as measured by time spent on these variables. Many techniques have been widespread in Brazil in order to monitor in afforestation, including: remote sensing and geoinformation systems. The use of these systems is scarce in the activities of arborization and requires research that highlight the benefits in the planning and implementation of urban forest inventories. This study aimed to implement the Geographic Information System (GIS) in the measurement of variables of the inventory the trees on the streets of Curitiba-PR, seeking agility and reducing costs of this service, to support planning and monitoring urban trees. For this research, we selected three sample units of street trees in the city of Curitiba from a tree inventory conducted in 1984. The criteria for selection of these samples were larger amount of trees and species diversity. The sampling units represented the following districts: Agua Verde, Bigorrilho e Jardim Social. The variables measured were: number of trees, canopy area and canopy spacing. The measurement of these spatial variables were processed using the program ArcGIS 9.2 orthophotos of sample units. These variables were statistically compared with the same variables obtained by conventional inventory (measured in field). We calculated the following indices of spatial arborization cover of tree index (CTI), tree density index (TDI), rate of sidewalk tree kilometer (RSTK) full occupancy index (FOI). Results showed that there was no statistically significant difference between the conventional method and GIS. The differences between the values obtained in the collection methods are related to the following factors limiting the detection: the quality of cartographic databases used, errors of observation and measurement in conventional inventories; the period and time of image acquisition, the height of buildings (shadows), the phenological aspect of the species, the planting of different sizes or size. CTI obtained from three samples showed lower value than the bibliography, indicating the need to plant new trees in the areas analyzed. The speed and efficiency of a GIS in the measurement of variables, it is considered that this technique can assist and complement the conventional inventory. It is noteworthy that this method does not replace the conventional method can support the implementation of inventory, reducing the variables measured in the field that have a high cost of measurement. By the results it is concluded that obtaining environment variables dendrometric in GIS is a method currently applicable to the characteristics of the trees on the streets of Curitiba, mainly for the ease and availability of municipal cartographic databases.
