Therefore, this will be our Cloud mask input. Each L2A product contains a complete set of band-type rasters including a Cloud mask processed with sen2cor tool. In the other hand, the Copernicus Sentinel-2 mission is ideal as data provider of Cloud masks. Of course, you can use your own layer, this is only my sample data. This layer contains polygons of several Landuse types, we will have to extract only Urban-type polygons later. This layer is available to download from the SDI of Navarra (IDENA): In this guide I will use a Landuse polygon layer of the region where I live in ( Navarra, in the north of Spain). A data provider that publish Cloud masks, or something to get them, that overlap our polygons and over the time.A Vector layer with our polygons of interest, or a Raster layer with them rasterized.To calculate the Cloud coverage of a set of Urban areas, we will need two main data sources: I am sure you will have in mind your data whose digesting could be reached in a similar way. :-( Yes, I know, it is not an innovative data, also, we can get it from a lot of other sources on internet, but I think the interesting thing here is the way that I have to walk to get this value. In other words, we are going to awnser, using Sentinel2 data and a serverless platform on the cloud, this question: how sunny are some cities over the time? Let's me keep the guide simple our derived data will be very simple, a scalar value for each tile processed, the Cloud coverage of Urban areas of a region. This post describes how you can keep up-to-date your own derived data from Sentinel-2 imagery archive using a serverless architecture on AWS.

Sunny europe catalogue 2017