Google Earth Engine tasks

The main aim of the task is to analyze the vast amount of remotely sensed datasets and to automate the process. To complete the task, you will need to have access to Google Earth Engine ([login to view URL]), and you need to have R installed. For each of the following five tasks, follow the steps, document the workflow. Keep in mind that for each of the tasks along with the report, you will have to submit the code as well (in case of EarthEngine include the link to the shared code, and in the case of R, attach the file). 1. Create a processing chain for Landsat 8 imagery in GEE. • Get the Landsat-8 Surface Reflectance collection, filter it by time and location. • Apply a cloud masking function on a collection* • Apply a function to calculate NDVI over a collection and calculate the maximum of the collection. 2. Binary classification (water/no water, forest no forest) • Get Sentinel-2 data, filter it a with date, location, and metadata (select images with cloud cover is less than 10%) and make a median composite. • Collect some training data for water/no water areas (you can choose another binary class such as forest/no forest, this depends on the area of interest selected and the landscapes in that area) and apply a classifier (e.g., Random Forest) for delineation of the areas. 3. Use the MODIS collection to inspect time series over different biomes. • Get MODIS NDVI collection (make sure to select the collection with 16-day frequency), apply filters and plot times series for different biomes. Describe the patterns in the series that you extracted. 4. In this section you will use BFAST on the NDVI time series that you extracted in EE (from 5 areas of interest) to detect changes within the time series. Use the time series from the previous task, read it in R and apply the BFAST function. Copy the resulting R BFAST graph in the report and describe the detected components and change types detected within the time series. 5. Get Land Surface Temperature and NDVI data from MODIS* and calculate the correlation between these two datasets.