Comparative analysis of three NDVI resolutions in different urban typologies. The case of Granollers in the Metropolitan Area of Barcelona.

In recent years, the use of remote sensed NDVI has become recurrent in urban studies regarding the adaptation of cities to climate change. However, due to the physical diversity within cities and the different resolution offered by the sensors, the territorial interpretation of what the NDVI values really mean becomes difficult. Where the larger the size of the cells of the image, the greater the number of elements of the built environment within it, and the more complex the interpretation becomes.

In this work, the relationship between the NDVI of three sensors with different cell resolution for the same location and date is studied. In particular, the city of Granollers in the Metropolitan Area of Barcelona is analyzed. First, the NDVI images were obtained from Landsat-8 with 30m resolution, Sentinel-2 with 10m and from the Ministry of Agriculture, Livestock, Fisheries and Food of Catalonia (DARP) with 0.125m resolution. Then, the comparison was performed with a sample of five different typologies of the territory: dense urban core, suburban, industrial, area of highway and rural.

As first results, a supervised classification of the DARP image allowed the definition of 0.30 as the precise minimum value of NDVI that indicates the actual presence of vegetation. On the other hand, the comparison indicates that, in the urban context, the larger the cell size, the presence of vegetation quality is overestimated, where the higher percentage of cells is concentrated in higher NDVI values than in those with lower resolution. However, this behavior is not appreciated in rural areas, where higher percentages of cells of different resolutions were concentrated in the same NDVI ranges.

In such a way, it is corroborated that it is in the urban context where this indicator has a greater difficulty of territorial interpretation. Statements that are analyzed in greater depth in this study, where its implications in the use of NDVI in urban studies for the adaptation of cities to climate change are discussed.