Plants, floods and dams: a model that predicts how rivers and their banks respond

Plants shape a river as much as the river shapes the plants, and a model tested in Dinosaur National Monument aims to predict both. The Northern Colorado Plateau Network, the U.S. Forest Service and the University of Montana built it to help park managers see how changes in river flow may affect downstream vegetation and riparian topography, and the other way round.

Response curves link a response variable (the Y axis) to control variables (the X axis). Image from the National Park Service's page.

How plants shape rivers

Dams stabilize a river's natural flow to produce electricity, store water and reduce flooding, and slow, steady flow with long inundation promotes water-loving (hydric) plants. In summer base flow, native willows and non-native tamarisk take hold low in the channel. Dense patches slow the water and trap sediment; as sediment builds, a stable base flow can encourage more growth, and an active channel can turn into a stable floodplain that grows higher and narrows the channel.

Eventually the floodplain can grow too high for native willows. Tamarisk, adapted to dry land, can thrive on high surfaces, and its deep roots hold soil and divert water, deepening the main channel. A gradual slope from channel to floodplain can become a steep bank armored by tamarisk, which makes it hard for boaters to reach campsites and harms conditions for fish.

Undammed rivers, and dammed ones managed to mimic natural variation, stay diverse by keeping plants "on their toes." Plants must grow fast after floods recede, root deep or flex under the water's force. Water knotweed, for example, grows as an aquatic plant in spring runoff and becomes a land plant as the water drops. Flows that vary from year to year stop any one type from dominating, and floods after snowy winters wash some plants out and open new habitat for others.

The model

A model was developed to predict how groups of plants ("guilds") respond to flow patterns, and it was tested at one site on the Green River, whose flow is partly regulated by Flaming Gorge Dam, and one on the unregulated Yampa River. It uses "response curves" showing how one variable, such as a plant guild, changes with another, such as flow. A curve's shape can reveal thresholds where a response is sensitive or resilient. Using vegetation data collected by the network, the model predicted where six guilds were most likely to occur given flow velocity and length of inundation, and how dense the vegetation of a plot would be. It also considered how vegetation changes flow and tried to predict how topography would change in a flood.

The model predicted the hydric herb guild especially well. Dots show observed presence (red) and absence (empty) at the Seacliff site on the Green River, with suitable habitat shaded blue. Image from the National Park Service's page.

What it found

Vegetation. The model accurately predicted the general distribution of guilds along the river. Hydric guilds were likeliest in plots flooded for most of the growing season, xeric guilds in dry places, and mesic guilds between the two. Faster water and longer flooding made plots less likely to be vegetated. Hydric and xeric guilds could occur across a wide range of velocities up to 4 meters per second, while mesic guilds favored low to moderate speeds. At plots with mature woody guilds, dense plant cover was 20 to 70 percent more likely than elsewhere.

Topography. The model had some success predicting erosion or deposition at a plot after a single flood, but it was better at predicting which plants occurred. The best models took in velocity, sediment load and vegetation. On the regulated Green River vegetation cover strongly determined the topographic response; on the Yampa, non-living factors predicted it better. Other studies suggest vegetation matters more when sediment load is out of balance with the river's capacity to carry it, as on the regulated Green.

Left: observed topographic response to the 2016 flood at the Seacliff site. Right: the model's prediction. Where the colors match, the prediction was accurate. Images from the National Park Service's page.

Why it matters

Flows on regulated rivers could be changed to make more power, reduce flooding or help endangered fish, and unregulated rivers could change through climate change or diversions. Changes in flow could favor or discourage stands of tamarisk. After years of stable flows and small peaks, for instance, vegetation might crowd habitat and favorite camps, and managers could ask for a high spring release to scour it. After removing tamarisk, they could ask for a bigger peak to move trapped sediment. Long-term monitoring will test the model, and feeding post-change data back in should improve its predictions.

Sources

• National Park Service, Northern Colorado Plateau Network, "Predicting Vegetation and Topographic Change in Response to Altered River Flows on the Northern Colorado Plateau" (https://www.nps.gov/articles/ncpn_flowandveg.htm), a brief summarizing R. M. Diehl, A. C. Wilcox, D. M. Merritt, D. W. Perkins and J. A. Scott, "Development of an eco-geomorphic modeling framework to evaluate riparian ecosystem response to flow-regime changes," Ecological Engineering 123, 2018; rewritten in hubnx's own words. The page's photograph of people at a river camp is left out.