Near real-time intercomparison of satellite and webcam derived snow cover

Webcam image processing for obtaining environmental data has emerged in latest years. Digital imagery from environmental camera networks, which are mostly available for free, can be used in such processing techniques. FMIPROT Camera Network Portal is a platform that brings data and metadata from multiple camera networks and uses them to provide near real-time observations for different parameters, e.g., vegetation indices, snow cover and snow depth. Snow cover is an input parameter for weather forecasts, climatology and hydrology; therefore, it is essential for assessing natural hazards such as avalanches or floods and managing associated risks.

Global Observing System for Climate (GCOS) has specified the snow cover as part of the 50 essential climate variables (ECVs) to be observed by satellite remote sensing. Copernicus is providing multiple satellite-derived snow cover products through Copernicus Land Monitoring Service. Our latest studies showed that snow cover data derived from webcam imagery can be used in the validation of satellite-derived products.

We have used the FMIPROT Camera Network Portal as a platform to intercompare the produced fractional snow cover data with the Copernicus snow cover products. Our aim is to assess the feasibility, benefits and drawbacks of using webcam imagery collected from different sources for near real-time validation of Copernicus products. This study is implemented in the ESA IDEAS+ project, as a part of the QA4EO framework.

First, a simple HTTP API is developed to be able to fetch the webcam derived data from the platform, according to the temporal and spatial coverage. This API is hosted in the portal, which is also available for free. Then, the API is used to visualize the latest observations from the webcam imagery on interactive maps in the portal webpage. Later, processing chains are established to download Copernicus products in the server and compare the observations using the API. The comparison chains are run on the server regularly, every day for the last 7 or 21 days of data and every month for the last 1 year of data. The results are stored on the server. RMSE values and data pairs are visualized on interactive maps and in scatter plots respectively.

The coverage of the webcam data is increased in the operational monitoring processing chains by adding more camera data. Now, data from 15 cameras in Finland from the MONIMET camera network, FMI weather stations and Finnish airports and 2 cameras in the Alps from the PhenoCam Network are included in the comparisons. The satellite products used in the comparisons are Copernicus Snow Cover Extent Europe (500m), Copernicus Snow Cover Extent Northern Hemisphere (1km) and Copernicus High-Resolution Snow Cover Pan European (20m). Our efforts continue to include more webcam and satellite data in the comparison processing chains, to implement two-dimensional location metadata in the processing software to include webcam data in complex terrain and to study different methods in the retrieval of the snow cover from webcam imagery.