July 06, 2020
Machine Learning-Based Zonation to Understand Snow, Plant, and Soil Moisture Dynamics Within a Mountain Ecosystem
Unsupervised learning help to identify spatiotemporal patterns of snow-plant dynamics in high-resolution time-lapse data.
The Science
In the headwater catchments of the Rocky Mountain region, plant dynamics are largely influenced by snow accumulation and melting as well as water availability. Key properties such as snow coverage, soil moisture and plant productivity are highly heterogeneous in mountainous terrain. This study identifies the spatiotemporal patterns in co-varied snow, plant, and soil moisture dynamics associated with microtopography based on high-resolution satellite imagery and unsupervised machine learning.
The Impact
Researchers found that unsupervised learning methods can reduce the dimensionality of timelapse images effectively. The results identify spatial regions—a group of pixels— that have similar snow-plant dynamics (based on Normalized Difference Vegetation Index) as well as their association with key topographic features and soil moisture. This cluster-based analysis can tractably analyze high-resolution timelapse images to examine plant-soil-snow interactions, guide sampling and sensor placements, and identify areas likely vulnerable to ecological change in the future.
Summary
In the headwater catchments of the Rocky Mountain region, plant productivity and its dynamics are largely influenced by water availability. Understanding and quantifying the interactions between snow, plants, and soil moisture has been challenging. These interactions are highly heterogeneous in mountainous terrain, particularly as they are influenced by microtopography within a hillslope. In this study, researchers investigated the relationships among topography, snowmelt, soil moisture, and plant dynamics in the East River watershed, Crested Butte, Colorado, based on a time series of 3-meter resolution PlanetScope Normalized Difference Vegetation Index (NDVI) images. To make use of a large volume of high-resolution timelapse images, researchers used unsupervised machine learning methods to identify the spatial zones that have characteristic NDVI time series and to reduce the dimensionality of time lapse images into spatial zones. Results show that identified zones are associated with snow-plant dynamics and microtopographic features. In addition, soil moisture probe and sensor data confirm that each zone has a unique soil moisture distribution. This cluster-based analysis can tractably analyze high-resolution timelapse images to examine plant-soil-snow interactions, guide sampling and sensor placements, and identify areas likely vulnerable to ecological change in the future.
Principal Investigator
Haruko Wainwright
Lawrence Berkeley National Laboratory
[email protected]
Program Manager
Jennifer Arrigo
U.S. Department of Energy, Biological and Environmental Research (SC-33)
Environmental System Science
[email protected]
Paul Bayer
U.S. Department of Energy, Biological and Environmental Research (SC-33)
Environmental System Science
[email protected]
Funding
This work was supported in part by the U.S. Department of Energy, Office of Science, Office of Workforce Development for Teachers and Scientists (WDTS) under the Science Undergraduate Laboratory Internship (SULI) program and Workforce Development and Education at Lawrence Berkeley National Laboratory. This material is based upon work supported as part of the Watershed Function Scientific Focus Area funded by the U.S. Department of Energy, Office of Science, Biological and Environmental Research Program under Award Number DE-AC02-05CH11231.
References
Devadoss, J., et al. "Remote Sensing-Informed Zonation for Understanding Snow, Plant, and Soil Moisture Dynamics within a Mountain Ecosystem." Remote Sensing 12 (17), 2733 (2020). https://doi.org/10.3390/rs12172733.