Monitoring freshwater ice phenology in Canada's North in the era of abundant open-access satellite sensors
Abstract
The ice phenology of freshwater lakes throughout the Northern Hemisphere has undergone important climate-induced shifts over the past century. In Canada's North, where freshwater lakes and wetlands cover 15 to 40% of the landscape and the duration of the ice-free season has been increasing, monitoring ice phenology is vital given its important impacts on climate, socio-economic, ecological and hydrological systems. Remote sensing techniques are particularly well suited to tracking phenological events (e.g.spring break-up and autumn freeze-up) over vast and remote regions. The rapid nature of freeze and thaw phenology events has restricted large-scale monitoring efforts to satellite sensors with frequent revisit times (e.g.MODIS, AVHRR), but these are limited to larger water bodies constituting only a small fraction of northern landscapes. However, the increased abundance of open-access satellite imagery from different sensors provides opportunities to reduce the trade-off between spatial and temporal resolution.Here we present a Google Earth Engine (GEE) algorithm that fuses information from multiple sensors into a coherent time series of freshwater ice phenology estimates over all of Canada at a 30-metre resolution for the spring seasons of 2014 to 2017. Using both optical (Sentinel 2 and Landsat 8) and SAR (Sentinel 1) imagery, we built reference datasets from a variety of lakes to create simple classification trees distinguishing ice/snow, water, and clouds. We combined the provisionally classified images into a single time series and built a pixel-wise logistic regression to remove misclassifications, using the resulting time series to estimate the date that each pixel transitioned from ice to water in a given year. The result is a weekly time series showing the breakup of ice within and among hundreds of thousands of lakes, which compared well with the Canadian Ice Service's database of weekly ice cover estimates for around 100 lakes. Capitalizing on the growing abundance of free open-access satellite imagery, our analysis provides: (i) an accurate estimate showing the dynamics of spring break-up events at a high spatial resolution, (ii) a scalable method readily applied to other subarctic and arctic regions, and (iii) a coherent time series easily improved by adding classified imagery from any sensor.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.H31K2072C
- Keywords:
-
- 1819 Geographic Information Systems (GIS);
- HYDROLOGYDE: 1855 Remote sensing;
- HYDROLOGYDE: 1856 River channels;
- HYDROLOGYDE: 1857 Reservoirs (surface);
- HYDROLOGY