Climate adaptation, Water conservation, Water resources
Developing Tools for Improved Water Supply Forecasting in the Rio Grande Headwaters
Case Study by Conservation and Adaptation Resources Toolbox and the Drought Learning Network
Status
Completed

Location

States

Colorado

Ecosystem

Montane, River/stream

Subject

Climate change
Drought
Fires
Forests
Hydrology
Invasive species
Invertebrates
Montane
Rivers and streams
Water budget
Watershed

Introduction

Barren trees in a forest
Image Details

The Rio Grande starts in Colorado’s San Juan Mountains, runs through New Mexico and along the New Mexico-Texas border, and flows into the Gulf of America. This river serves as an important natural resource for community use, providing water for irrigation for 2 million acres of land, drinking water for over six million people, hydroelectricity, and opportunities for recreation. 

The Rio Grande Compact of 1939 (RGCC 1939) governs the streamflow across the Colorado-New Mexico border and the New Mexico-Texas-Mexico border. Water resource managers rely on several decision support tools that inform the appropriate diversion, retention, and release of streamflow that meet measures mandated by the RGCC and are required for thriving communities. The Natural Resource Conservation Service (NRCS) Water Supply Outlook (WSO) reports serve as useful tools for water management by providing streamflow runoff predictions during the snowmelt runoff season. 

Snowmelt from the mountains of Colorado contributes to more than half of the river’s streamflow, making it a valuable source of water. However, models that inform WSO reports do not consistently account for factors that might influence changes in snowpack, such as rising temperatures or decreasing tree canopy density. Given recent changes in climate and land cover in the Upper Rio Grande (URG) region, streamflow forecasting could be less reliable than in the past. Streamflow models that are unable to account for recent changes in temperatures and land cover compound uncertainty in planning and allocating water for reliant downstream communities. 

To address this concern, researchers from the U. S. Geological Survey (USGS) Colorado Water Science Center sought to identify potential sources of errors and improve forecasting for the URG region. Improved forecasting will equip stakeholders with the tools to plan water use effectively and inform efforts in other regions where declining snowmelt poses a severe threat to community water supplies.

Key Issues Addressed

NRCS WSO reports serve as a common decision-making tool among water resource managers in the URG region. These reports are informed by a network of snow telemetry stations (SNOTELs) which are used to predict the volume of streamflow during the runoff season based on statistical relationships between snow, precipitation, streamflow, and forecaster judgment. 

However, the statistical models used by SNOTELs are limited and do not account for many outside environmental factors that can affect relationships between model variables. For example, climate and land cover have changed in the URG region, an area susceptible to drought, bark beetle infestation, and wildland fire. The combination of these factors can influence the water cycle and energy balance in the ecosystem. For instance, beetle infestation kills trees and reduces canopy cover, which allows more sun exposure and increased evaporation from snowpack into the atmosphere. Moreover, the relative combined influence of climate change climate change
Climate change includes both global warming driven by human-induced emissions of greenhouse gases and the resulting large-scale shifts in weather patterns. Though there have been previous periods of climatic change, since the mid-20th century humans have had an unprecedented impact on Earth's climate system and caused change on a global scale.

Learn more about climate change , beetles, and wildland fire on streamflow forecasts are unknown. 

Without consideration for these variables, streamflow forecasting tools can become increasingly unreliable. Inaccurate predictions of streamflow availability could cost millions of dollars due to ill-informed allocation of water and loss of agricultural productivity. As climate and land cover change grow in occurrence and severity across the globe, the question of streamflow forecast reliability is not unique to the URG region. Efforts to address streamflow forecast reliability in this area may contribute to successful water management in similar communities worldwide.

Project Goals

  • Identifying Forecast Errors: Suspicions of declining forecasting accuracy have developed in Colorado and New Mexico over the last decade but have yet to be confirmed. The research team aimed to investigate this potential decline in forecast reliability by testing for prediction bias trends of WSOs from 1990 to 2017. Further, the research team sought to identify factors that contribute to prediction bias and their magnitude of influence by incorporating climate and land cover change measures. 
  • Improving Snow Measures: SNOTELs provide point-based measurements of snowpack for developing WSOs, which risks ignoring basin-wide changes in snowpack characteristics from novel environmental changes. Recent climate and land cover changes may alter basin-wide snowpack characteristics to an extent that streamflow is altered as well. By comparing SNOTEL outputs to simulated basin-wide snow water equivalent (SWE) and snow-covered area (SCA) from 1984-2017, the research team sought to calibrate for measurement differences to represent snowpack more accurately in current models. 
  • Modeling Environmental Influences: To evaluate the effects of continuous climate and land cover change on hydrologic systems in the URG region, the research team performed numerical experiments with physically-based hydrologic models. Models simulated the loss of forest canopy cover resulting from beetle infestation and wildland fire, accounting for current precipitation, air temperature, and solar radiation conditions.

Project Highlights

Reliability Decline: Declining forecasting reliability was confirmed by comparing periods of over- and under-prediction. Changing conditions, particularly fire’s influence on land cover, significantly influences forecasting reliability.

  • Shifting Snowmelt Season: Changing climate in the URG region contributes to a shift in runoff seasonality. Warming springtime temperatures matched with decreasing springtime precipitation result in less snow accumulation and earlier runoff in the spring. These factors probably contribute to prediction bias in streamflow forecasts and over-predicting seasonal snowmelt runoff.
  • Wildland Fire Influence: Bark beetle-induced tree mortality did not significantly influence streamflow in bark beetle-infested forests. However, land cover change as a result of wildland fire had a strong positive influence on seasonal runoff forecasts, increasing seasonal streamflow by an average of 28% to 35% for four years after wildland fire occurrence. The hydrologic effect differences between bark beetles and fire may be a result of the tree mortality patterns. While wildland fire inflicts larger scale tree mortality, bark beetle damage manifests in patches throughout the affected forest. This finding suggests that land cover is an important factor to consider to improve forecasting accuracy.
  • Future Precipitation Importance: Although land cover change, climate, and better-represented snow measures explained much of the forecast variance, a substantial source of variance remained unidentified. Through further numerical experimentation, the research team found that future precipitation projections for spring and summer months could help further improve forecasting. The National Weather Service Climate Prediction Center provides 30- or 90-day precipitation outlooks for springtime that can be used to inform early season streamflow forecasts. 
  • Visualization Tool: The research team created a simple spreadsheet tool to aid in visualizing the importance of land cover change and climate on seasonal streamflow. The user can input values for SWE, acres burned, or precipitation, and the chart will update with new streamflow predictions. While not a suitable replacement for NRCS streamflow forecasts, this tool allows the user to understand the relative importance of land cover and climate on these forecasts.

Lessons Learned

Mountains covered in forests
Image Details

The research team approached the issue from a system-wide perspective. This holistic approach, which explored multiple dimensions of streamflow with a variety of research methods, allowed the research team to gain more information and greater clarity for envisioning solutions. 

Managing data presented a large, but achievable, challenge for the research team. Exploring multiple aspects of streamflow with a variety of research methods required attention to detail in the data compilation and organization process. 

Predicting summer precipitation is even more challenging than springtime precipitation due to precipitation variability. However, understanding summer precipitation should not be ignored, especially in regions like the URG that receive both substantial winter snow and summer rain from the North American Monsoon. Efforts to improve precipitation forecasting will further advance streamflow forecasting. 

Increasing springtime air temperature and decreased snow accumulation has resulted in an earlier snowmelt season. As air temperature and snowpack could continue to change, it is important to monitor the occurrence and degree of its influence on snowmelt to maintain accurate streamflow forecasts.

Next Steps

  • Snowpack duration is changing; however, the influence this phenomenon has on fire season is unknown. Building off of this project, the researchers aim to determine the effect that shifting snowpack seasons has on the duration and severity of fire season.
  • With a better understanding of bark beetle and fire influence on runoff quantity, the research team will begin to explore these factors’ relationship with water quality.
  • The project results provide critical information for the URG region, and other areas whose water originates from snowpack. For effective water conservation and distribution, it is imperative to bridge the gap between the research results and monitoring, management, and policy. The research team aims to assist practitioners in applying results to real-world scenarios.

Funding Partner

South Central Climate Adaptation Science Center

Resources

Contacts

CART Lead Author

  • Maude Dinan, Program Specialist, USDA Southwest Climate Hub: mdinan@nmsu.edu

The DLN is a peer-to-peer knowledge exchange between climate service providers and resource managers, created to gather and share lessons learned from drought events to prepare for future events. The DLN partners with CART to develop Case Studies, with funding from the National Drought Mitigation Center for interns and coordination support from the USDA Southwest Climate Hub.

Suggested Citation

Dinan, M., E. (2021). “Developing Tools for Improved Water Supply Forecasting in the Rio Grande Headwaters.” CART. Retrieved from https://www.fws.gov/project/tools-water-supply-forecasting-rio-grande.

Programs

The Conservation and Adaptation Resources Toolbox logo which includes a butterfly flying over a stream with a fish in it. On the stream bank there are two trees and a windmill.
For over eight years, CART enhanced collaborative conservation efforts at all scales by facilitating issue-based, not geography-based, peer-to-peer knowledge sharing. By connecting hundreds of individuals from dozens of organizations across North America, CART helped bridge the gaps between work at...
Program

Partners