Contents lists available at ScienceDirect Urban Forestry & Urban Greening journal homepage: www.elsevier.com/locate/ufug Special article Understanding multi-temporal urban forest cover using high resolution images Aline Canetti a , Marilice Cordeiro Garrastazu b , Patrícia Póvoa de Mattos b, ⁎ , Evaldo Muñoz Braz b , Sylvio Pellico Netto a a Federal University of Paraná, Av. Pref. Lothario Meissner 900, 80210-170, Curitiba, PR, Brazil b Embrapa Florestas, Estrada da Ribeira Km 111, Caixa Postal 319, 83411-000, Colombo, PR, Brazil ARTICLE INFO Keywords: Urban forestry Image processing Object-based classification ABSTRACT Urban forests offer city residents a better quality of life. They also act as barrier and filter to pollutants resulting from human activities. Mapping the distribution of forests in urban areas is therefore important for good urban planning. Currently, high resolution images are considered useful tools to quantify forested urban areas for large scale urban development. The objective of this work was to evaluate multi-temporal urban forest area changes over different land-use classes, using available high-resolution images obtained from different satellites. The municipality of Araucaria (Paraná State) was chosen as study area because it is a large industrial zone of southern Brazil. The effects on the forests within this municipality caused by the increase in human activities between 2005 and 2012 were determined using high resolution images (5 m). We recorded a reduction of 22.8% in the forests surrounding urban areas of the municipality, as a result of deforestation of 791 ha and plantation of only 251 ha. The utility, commercial and residential zones which are more crowded the areas of highest po- pulation density were those which showed the greatest loss of tree cover. Object-based classification accuracy using images from different satellites was sufficient to quantify the evolution of tree-cover over the studied period. 1. Introduction There is a growing recognition that urban forests improve urban quality of life in many ways, offering benefits that meet local needs (Küchelmeister, 2000). Through the physiological processes of trees, forests act as air filters and have a direct influence on temperature and other climatic variables involved in air pollutants dispersion (Murphy et al., 1977; Lima, 1980; Brack, 2002; Walton et al., 2008). Preservation of forests in urban and industrial areas is therefore of fundamental importance. Generally, industrial areas are associated with environmental impact, such as smoke, construction and noise. Predomination of asphalt and concrete in those areas increases summer heat by raising air and soil temperature which consequently reduces moisture in tropical and subtropical regions. The limited space of ex- posed soil makes it more difficult for precipitation and air to percolate into soil, which reduces nutrients necessary for plant growth. This in- creases as soils become compacted by the transit of heavy vehicles and heavy machinery damages plant roots, trunks and branches (Poracsky and Scott, 1999). The role of ecosystems has become especially relevant with in- creasing human impact and environmental pollution. The environ- mental and economic values of forests and their sensitivity to pollution make them indicators of environmental changes, which allow model- ling general tendencies of the whole biosphere affected by human ac- tivity (Juknys et al., 2002). To predict consequences of a marked increase in urbanization on quality of life requires an understanding of the evolution over time of a number of variables. For example it is important to understand the relationship between the expansion of urban areas and the distribution of remaining forest cover. One way of modelling changes and their consequences is to study variations in municipal green areas at a global level, examining different urban areas in various parts of the world (Fuller and Gaston, 2009). Tree cover may be mapped by different methods when using images. Some authors use remote sensing for this purpose (Myeong et al., 2001; Sawaya et al., 2003; Walton et al., 2008; Buccheri Filho and Nucci, 2011; Myint et al., 2011; Eduardo et al., 2013; Nagi, 2014) and others field inventories (Banks et al., 1999; Harder et al., 2006). https://doi.org/10.1016/j.ufug.2017.10.020 Received 30 May 2017; Received in revised form 23 October 2017; Accepted 29 October 2017 ⁎ Corresponding author. E-mail addresses: alinecanetti@ufpr.br (A. Canetti), marilice.garrastazu@embrapa.br (M.C. Garrastazu), patricia.mattos@embrapa.br (P.P.d. Mattos), evaldo.braz@embrapa.br (E.M. Braz), sylviopelliconetto@gmail.com (S. Pellico Netto). Urban Forestry & Urban Greening 29 (2018) 106–112 Available online 11 November 2017 1618-8667/ © 2017 Elsevier GmbH. All rights reserved. MARK