Quantifying Surface Water Dynamics at 30 Meter Spatial Resolution in the North American High Northern Latitudes 1991–2011

Abstract

The availability of a dense time series of satellite observations at moderate (30 m) spatial resolution is enabling unprecedented opportunities for understanding ecosystems around the world. A time series of data from Landsat was used to generate a series of three maps at decadal time step to show how surface water has changed from 1991 to 2011 in the high northern latitudes of North America. Previous attempts to characterize the change in surface water in this region have been limited in either spatial or temporal resolution, or both. This series of maps was generated for the NASA Arctic and Boreal Vulnerability Experiment (ABoVE), which began in fall 2015. These maps show a "nominal" extent of surface water by using multiple observations to make a single map for each time step. This increases the confidence that any detected changes are related to climate or ecosystem changes not simply caused by short duration weather events such as flood or drought. The methods and comparison to other contemporary maps of the region are presented here. Initial verification results indicate 96% producer accuracy and 54% user accuracy when compared to 2-m resolution WorldView-2 data. All water bodies that were omitted were one Landsat pixel or smaller, hence below detection limits of the instrument.

Multi-Decadal Surface Water Dynamics in North American Tundra

“…The methodology used to create the annual water maps from a series of water classifications is robust and has been employed for both Landsat and MODIS instruments [22,28] and is similar to the method adopted for the Global Landsat Water maps [27]. The development of these new time series products of surface water make it possible to apply our object based analysis methodology in other regions.…”

“…To ensure that there are enough observations for the maps to depict nominal (neither maximum nor minimum) extent, a rolling set of cloud free observations over three years of inputs are used to create a single map (e.g., data from 1984, 1985, 1986 used for 1985 map; data from 1985, 1986, 1987 used for 1986 map; etc.). This approach was used effectively with Landsat data to generate the decadal water maps in the high northern latitudes [28].…”

“…Water in the annual map is anything that has greater than or equal to 50% probability of being water. Following this methodology, annual maps were generated for each year from 1985 to 2015; full description can be found in [28].…”

“…Partial water A combination of high and low confidence water was processed through the same post classification procedure (described below) that was used to generate the decadal water maps [28] to produce annual maps ( Figure 2). Though the partial water class correctly identifies edges of water features it also seems to identify many false positives and hence was treated as "not water" in the annual map creation.…”

“…Each of these products has represented an advance in the mapping capabilities but also still has significant limitations including spatial resolution [22][23][24], temporal resolution [25,26] or both [27]. Most recently, work has been done to generate continental scale maps of surface water extent at the decadal time step using a time series of Landsat data as input [28,29]. Unlike many other 30 m spatial resolution water maps, the decadal water maps use three years of inputs to generate a single map in order to have enough repeated observations throughout the study region to have confidence that the water that was detected was not a false detection caused by cloud shadows or burn scars, and did not represent a local minimum or maximum caused by short-term weather events (i.e., drought or flood).…”