Glaciers & Students: the new glaciers inventory of Pakistan

This study has been developed in the framework of the UNDP (United Nations Development Programme) and EvK2CNR-Italy Project “Glaciers and Students - A scientific based approach to monitor climate and glaciers in Pakistan Mountain Regions to support hydrogeological risk prevention”. The main expected result of this project will be the updated glaciers inventory of Pakistan. The region of Hindu-Kush Karakorum ranges in Pakistan counts a number of more than 7,000 glaciers, and a large number of them are debris-covered glaciers. The improvement of knowledge of the cryosphere processes and the impacts of the climate changes can generate a new awareness in the local communities in formulating long term strategies for disaster risk reduction and for the environment sustainability. As resulting effects, better decisions and actions in these sectors will have positive impacts on the management of water and land resources, ultimately enabling to establish better responses to climate change and contribute in poverty reduction (Sustainable Development Goals No. 1, 6, 13 and 15). Starting from these expected outputs, and considering the new technical performance that can be activate in the using of satellite data, the strategy of this project can be divided into two main goals: Developing a new methodological approach to map the glaciers; Involving the Pakistani and Italian students into the process of recognition of the glacier’s boundaries. Preliminary from these purposes, the results proposed in this study are focused on the description of the discussed new methodology for glacier inventory. As well-known, several software can be used for image processing and editing in GIS environment, either free or commercial. In this project the IMPACT Toolbox (http://forobs.jrc.ec.europa.eu/products/software) has been chosen for the pre-processing and the semi-automatic classification of the satellite data. The selection and preliminary processing of the Sentinel-2 L2A images was carried out using the Google Earth Engine cloud-processing platform, on the basis of the “COPERNICUS/S2_CLOUD_PROBABILITY” collection, based on the “sentinel2-cloud- detector”, containing the probability (0-100) of each pixel of each Sentinel-2 image to be cloudy (or extremely clear). This information allows to filter the cloudy pixels present in the level 2A images contained in the COPERNICUS/S2_RS collection to facilitate the creation of cloudless composite. From the synthetic image of Sentinel 2, spectral indexes referred to snow and ice cover, and the image segmentation to partitioning of the composite into relatively homogeneous regions were applied. These regions (polygons) can belong or not to a glacier, and the interpretation by the user is supported by their boundaries. The final product is a map in vector format (shapefile), where each glacier results from the merging of the polygons inside its boundary. This method, and the software infrastructure planned in the study, has been considered suited to familiarize with the mapping process using satellite images by the students from the Italian and Pakistan universities involved in the project with a basic knowledge in remote sensing, GIS, and glacier’s mapping. For these topics, training activities have been organized and guidance materials in form of videos have been shared.