Alan Kasprak - Current Research

Improving Geospatial Techniques for Monitoring and Assessing Watershed Change

In addition to my research on assessing the impact of dams and reservoirs on rivers, I have additional interests in developing GIS tools and techniques for understanding how landscapes are changing through time. This work falls into two main areas:

  • Designing software to rapidly interpret the processes driving landscape change from geospatial data
  • Understanding the effect of topographic data resolution on our interpretation of landscape change

This page describes each of these projects in more detail.

Designing software to rapidly interpret the processes driving landscape change from geospatial data

At the moment, landscape scientists are faced with a good problem to have, but a problem nonetheless: we’re able to collect more topographic data, at higher resolutions, than we can ever hope to analyze in a timely manner. The advent of new technologies like terrestrial laser scanners and structure-from-motion photogrammetry mean that we can capture the form of landscapes at channel reach scales, but because of the time it takes to analyze those data, our ability to make fundamental statements about landscape evolution and response to human activities like land use, agricultural development, or river regulation is quite limited.

To address this, I’ve been working with Josh Caster and Joel Sankey (both at the USGS), and Sara Bangen (at Utah State University) to develop open-source software for rapidly and objectively discerning the sediment transport mechanisms that drive topographic change. These mechanisms can include pathways like fluvial or alluvial transport, mass failure, and aeolian (i.e., windblown) sediment transport. Using the software we’ve developed, we’re able to analyze repeat topographic datasets as they’re collected and quantify the relative role of sediment transport pathways in driving topographic change, in the process uncovering how those pathways change in response to anthropogenic landscape alteration or river regulation.

Inferred mechanisms of geomorphic change at a site along the Colorado River. Mechanisms were computed using original software for rapid and obective analysis of repeat topographic data.

In 2017, we published a manuscript in Earth Surface Processes and Landforms that detailed the development of this Python-based software. Check out that paper here! You can also download the software and example data from the USGS. This software is currently enabling stakeholders in Grand Canyon to understand the implications of shifting flow regimes from Glen Canyon Dam in terms of driving landscape changes throughout this iconic river corrior.

Understanding the effect of topographic data resolution on our interpretation of landscape change

Repeat topographic data can tell us a lot about how landscapes are changing through time. The usual approach is to begin by collecting information on the topography of an area of interest, commonly termed a 'digital elevation moel', or DEM. If we repeat this process some time down the road and create a second DEM, we can compare the 'before' and 'after' DEMs to quantify the amount of erosion or deposition ocurring across an area in the inter-survey period.

An example of sediment erosion and deposition, as inferred from the difference between two successive DEMs captured two years apart at a site in Grand Canyon.

With all the topographic surveying tools available to geospatial scientists, we can obtain DEMs at many different resolutions. For example, USGS DEMs constructed from topographic maps might have cell sizes, or resolutions, of about 10 meters. Airborne lidar DEMs are often 1 m in resolution, and ground-based lidar can provide DEM resolutions down to several centimeters. When it comes to using these DEMs for quantifying the erosion or deposition occuring across a landscape, should we always obtain the highest possible resolution of DEM, even though that often leads to bigger datasets and much longer processing times?

In collaboration with Nat Bransky, Josh Caster, and Teki Sankey (Northern Arizona University), along with Joel Sankey (USGS), I've been analyzing how much information on landscape change we gain, or lose, as DEM resolution varies.

It turns out that as we coarsen DEM resolution, we tend to dampen the amount of erosion or deposition occurring across an area of interest. Practically speaking, this means that for landscapes that are right at the line between gaining or losing sediment (something we call sediment equilibrium), the highest possible DEM resolution is important, because we might artificially lessen the inferred amount of erosion or deposition happening there. On the other hand, if we know a landscape is trending in one direction or the other - losing or gaining large volumes of sediment - then coarser DEM resolutions can be used, because the overall trends in the sediment budget won't change.

The amount of sediment being eroded (red) and deposited (blue), expressed as both the volume of change and the area of change, for four sites within the Grand Canyon. Note that as resolution coarsens, the amount of erosion and deposition are decreased, but deposition is affected more than erosion.

One of the other things we've found is that the way we interpret the mechanisms driving sediment movement vary based on DEM resolution. We used GIS software that we've been developing over the past several years to automatically interpret the processes driving landscape change from repeat topographic data. You can find information on that software at the top of this page.

Inferred mechanisms of geomorphic change at a site along the Colorado River across five DEM resolutions. Note that as resolution becomes coarser, the amount of change attributed to aeolian processes increases, whereas fluvial, alluvial, and colluvial processes decrease in importance.

As DEM resolution coarsens, we tend to attribute more changes in the landscape to aeolian processes, or those caused by wind. Conversely, with coarser resolutions, we see less sediment movement driven by fluvial (river), alluvial (gullying), or colluvial (mass wasting) processes. We anticipate that the results of this work will be helpful for scientists and land managers seeking to use repeat topographic data to determine whether, and how, sediment is moving through landscapes.

In 2019, we published a manuscript in Geomorphology on our analysis of DEM resolution and sediment movement; you can read that paper here.