Looking ahead to monitoring habitat changes in 2022
Hello again! Well, 2021 is almost over, so it seems like a good time to take a quick look back on what we’ve accomplished in the Colorado River basin and chart a course for what’s to come in 2022.
What we’ve accomplished
As a reminder, our project goal has been to connect vegetation trends over the last 35 years to watershed management and restoration practices. With our data, watershed managers can track vegetation trends through time and explore how those trends vary across a watershed. Our hope is that watershed managers will be able to identify which particular restoration techniques generate the landscape responses that they want, so they can make informed decisions in the future when targeting efforts and resources. On a larger scale, we hope to develop our techniques for detecting vegetation and land use change across a variety of landscapes to help researchers, conservationists, rangers and others detect potential problem areas in wild lands and protected spaces.
In June, we released our 35-year time series dataset for the Colorado River basin using Landsat satellite imagery. The data set includes seven different spectral indices that help characterize the habitat in an area: the Normalized Difference Vegetation Index (NDVI), the Enhanced Vegetation Index (EVI), the Soil Adjusted Vegetation Index (SAVI), the Modified Soil Adjusted Vegetation Index (MSAVI), the Normalized Difference Moisture Index (NDMI), the Normalized Burn Ratio (NBR), and the Normalized Burn Ratio 2 (NBR2). You can download the 35-year data set from this GitHub repo.
And in November, we released our Verde River map — a demonstration of what’s possible focused on a tributary of the Colorado River in Arizona, by combining our 35-year data set with high-resolution satellite imagery basemaps from Planet. In addition to all of the usual features of SkyTruth Alerts, users can access and view three new data sets in our Verde River map:
- a map of the Verde River and its subwatersheds with annual vegetation data for each subwatershed for 35 years (1985 through 2020);
- a collection of 35 Landsat-derived satellite image composites representing true color images of the complete watershed for each year, and;
- a collection of 35-year vegetation trend images of the watershed in seven spectral indices derived from Landsat.
The example above shows the trend in vegetation coverage and health of the Verde River area for every single 30-meter Landsat pixel. In this example, the pixels shown in deep orange are areas that have experienced statistically significant vegetation loss over the last 35 years, while pixels shown in deep blue are areas that have experienced statistically significant vegetation gain over the same period.
One important question is whether those vegetation gains reflect the spread of invasive species, such as tamarisk, that soak up a lot of water, lower the water table, and thereby exacerbate wildfires; or if they indicate native species returning to a previously degraded site. Clearly, this is essential information for our restoration partners who are working in these watersheds. Many locations are remote, difficult to reach stretches of river, and getting a restoration crew out there to remove tamarisk, russian olive, and other invasives is a huge undertaking – especially for small watershed groups with limited time, capacity, and resources. While our model can’t distinguish among species at this time, our approach will alert managers to specific areas of change that they may want to examine in more detail — through assessing high-resolution satellite imagery (read on!) or by conducting targeted field work.
We’re using SkyTruth Alerts and a new partnership with The Commons to bring our advanced analytics to watershed users where they need them, improving their ability to identify areas that they want to prioritize for restoration. And using new tools that we’re building with The Commons, watershed groups can more effectively and efficiently monitor and manage the huge landscapes that they steward going forward.
What to expect in 2022
We’re particularly excited about adding daily high-resolution satellite imagery from Planet Labs to map and track land use, vegetation cover, and habitat change in the Colorado basin in near real-time. Below you can see an example from just outside of Las Vegas in southern Nevada, where we’ve compared monthly satellite image composites from Planet to identify just over six acres of newly constructed, industrial-scale solar panels.
We are also excited about developing a machine learning approach to vegetation change in 2022. Adopting a machine learning approach will allow us and our partners on the ground to interpret vastly more imagery – by teaching computers how to analyze the ever-growing archives of satellite images for environmental changes of concern. And this is the vision for what we want to accomplish: a machine-learning powered land use change detection platform that we can apply not only to the Colorado River basin, but to protected areas across the globe that are important for restoring habitats essential for biodiversity and for minimizing and mitigating the impacts of a changing climate.
Make sure to check back in with us in 2022 for updates on our land use, land cover, and change-detection work. There are a lot of really exciting things happening in conservation technology, and we’re working hard to push the envelope. See you next year!