Characterizing Uncertainties in Climate Projections to Support Regional Decision-Making

Funding Amount and Duration:

$94,379 from April 1, 2016 - March 31, 2017

Funding Source:

US Geological Survey

Principal Investigators:

Adrienne Wootten, University of Oklahoma

About:

Global Climate Models (GCMs) use our understanding of atmospheric physics and other earth processes to simulate potential future changes in climate on a global scale. However, these large scale models are not fit for predicting smaller scale, local changes. Downscaling methods can be applied to the outputs of GCMs to give guidance appropriate for a more regional level. No standard approach to downscaling currently exists, however, and the process often results in climate projections that suggest a wide array of possible futures. It is critical that decision-makers looking to incorporate climate information understand the uncertainties associated with different downscaling approaches and can evaluate downscaled data to determine which datasets are appropriate for addressing their questions.

The goal of this project is to provide decision-makers with this information by evaluating the uncertainties associated with different downscaled datasets. Materials will then be developed to communicate these uncertainties to managers and explore how they can be incorporated into risk decision-making. The results will enable managers across the country to better understand possible climate futures in their jurisdictions, allowing them to make more informed planning decisions in the face of uncertainty.

Developing and Analyzing Statistically Downscaled Climate Projections for the South Central U.S.

Funding Amount and Duration:

$85,000 from April 1, 2016 - August 1, 2018

Funding Source:

US Geological Survey

Principal Investigators:

Adrienne Wootten, University of Oklahoma

Cooperators & Partners:

  • Berrien Moore III & Renee McPherson (OU)
  • Keith W. Dixon & John Lanzante (NOAA-Geophysical Fluid Dynamics Lab)

About:

Global climate models (GCMs) are a tool used to model historical climate and project future conditions. In order to apply these global-scale datasets to answer local- and regional-scale climate questions, GCMs undergo a process known as “downscaling”. Since there are many different approaches to downscaling there associated sources of uncertainty; however, downscaled data can be highly valuable for management decision-making if used with a knowledge of its limitations and appropriate applications.

In order to use downscaled data appropriately, scientists and managers need to understand how the climate projections made by various downscaling methods are affected by uncertainties in the climate system (such as greenhouse gas emissions and observed data). This project will produce 243 climate projections using three different downscaling methods, giving researchers insight into how each of these methods responds to various sources of climate uncertainty. This analysis will allow researchers to assist managers in selecting the best downscaled data for their specific management questions. This project will also result in foundational downscaled climate projections for the South Central region, assisting stakeholders in identifying the potential impacts of climate on a range of systems, from water to ecosystems to agriculture.

Informing Hydrologic Planning in the Red River Valley Through Improved Regional Climate Projections

Funding Amount and Duration:

$62,698 from September 26, 2015 - September 25, 2017

Funding Source:

US Geological Survey

Principal Investigators:

Ming Xue, Center for Analysis and Prediction of Storms (CAPS), University of Oklahoma

Cooperators & Partners:

  • Douglas Lilly (Co-PI) & David Williams (Co-PI), U.S. Army of Corps of Engineers (USACE)
  • Xiaoming Hu (Co-PI) & Renee McPherson, University of Oklahoma

About:

Across the Southern Great Plains, increasing temperatures are expected to alter the hydrological functioning of the region by contributing to severe droughts, more intense rainfall events, and more severe flooding episodes. These changes could adversely affect human and ecological communities. The ability to better predict future changes in precipitation and the response of hydrologic systems in the region could help mitigate their negative impacts. Yet while today’s global climate models provide large-scale projections of future temperature and precipitation patterns that can be broadly useful for large-scale water resource planning, they are often not appropriate for use at a smaller, more local scale.

This research will develop high-resolution climate projections for the Southern Great Plains that are better suited to informing water management at the local scale, with a focus on the Red River Valley. High resolution weather models will be used to downscale global climate model forecasts to provide more accurate local projections of future climate conditions for the Valley. These models will be run multiple times, creating a spread of model outcomes that will provide insight into the range of possible climate futures for the region and reveal any uncertainties managers should be aware of when using the projections. The very high-resolution projections will be used in the context of long-term hydrological modeling and management to inform cost-effective flood control planning, water supply management, hydroelectric power generation, and ecosystem conservation.

Quantifying Future Precipitation in the South Central U.S. for Water Resources Planning

Funding Amount and Duration:

$62,698 from September 26, 2015 - September 25, 2017

Funding Source:

US Geological Survey

Principal Investigators:

Jung-Hee Ryu, Texas Tech University (TTU)

Cooperators & Partners:

  • Katharine Hayhoe & Sharmistha Swain, Climate Science Center, Texas Tech University
  • Barry Keim, Kevin Robbins, Luigi Romolo & Amanda Lewis, Southern Climate Impacts Planning Program
  • Southern Regional Climate Center, Louisiana State University

About:

Publication: Observed and CMIP5 Modeled Influence of Large-Scale Circulation on Summer Precipitation and Drought in the South Central United States

The South Central U.S. is home to diverse climates and ecosystems, strong agricultural and energy sectors, and fast-growing urban areas. All share a critical need for water, which is becoming an increasingly scarce resource across the region as aquifers are overdrawn and populations grow. Understanding what brings rain to this region, and how the timing and amount of precipitation may be affected by climate change, is essential for effective water planning and management, yet community planners and managers have indicated that currently available precipitation forecasts for the South Central are insufficient, due largely to the high levels of uncertainty associated with precipitation projections for the region.

This project aims to improve scientific understanding of the local and large-scale atmospheric processes that bring moisture to the region and drive precipitation. The project will analyze long-term historical weather station records and atmospheric dynamics, improving our ability to interpret global climate model simulations and apply them to regional management questions. Researchers will project future changes in seasonal rainfall and drought risk to assist water resources planning and preparedness efforts.

Lessons learned from this work will be used to inform long-term projections for our region, making complex climate information and analyses more approachable, understandable, and actionable for regional policy-makers, planners, and managers.

Improving Representation of Extreme Precipitation Events in Regional Climate Models

Funding Amount and Duration:

$83,398 from October 1, 2013 - July 31, 2014

Funding Source:

  • U.S. Geological Survey, University of Oklahoma

Principal Investigators:

  • Ming Xue, University of Oklahoma

About:

Publication: An evaluation of dynamical downscaling of Central Plains summer precipitation using a WRF-based regional climate model at a convection-permitting 4 km resolution

Final Report

The South Central U.S. encompasses a wide range of ecosystem types and precipitation patterns. Average annual precipitation is less than 10 inches in northwest New Mexico but can exceed 60 inches further east in Louisiana. Much of the region relies on warm-season convective precipitation – that is, highly localized brief but intense periods of rainfall that are common in the summer. This type of precipitation is a significant driver of climate and ecosystem function in the region, but it is also notoriously difficult to predict since it occurs at such small spatial and temporal scales. While global climate models are helpful for understanding and predicting large-scale precipitation trends, they often do not capture many of the smaller atmospheric and earth surface processes that influence local and regional precipitation trends, like convective precipitation.

To address this gap in climate modeling capabilities, researchers developed regional climate models that are better able to project small-scale precipitation patterns and localized extreme precipitation events. Researchers combined information about land surface and water conditions with weather and climate models in order to quantify the local-scale impacts of climate on water resources. This highly localized information will assist regional decision-makers in addressing the challenge of predicting precipitation in the South Central U.S., leading to a better understanding of potential future impacts on agriculture, fish and wildlife, water quality and availability, and cultural resources.

Testing Downscaled Climate Projections: Is Past Performance an Indicator of Future Accuracy?

Funding Amount and Duration:

$124,393 from September 29, 2013 - September 24, 2016

Funding Source:

US Geological Survey

Principal Investigators:

John Lanzante, NOAA Geophysical Fluid Dynamics Lab

Cooperators & Partners:

Anne Stoner (Co-PI), Texas Tech University

Keith Dixon (Co-PI), NOAA Geophysical Fluid Dynamics Lab

Venkatramani Balaji (Co-PI), Princeton University

About:

Publication: Evaluating the stationarity assumption in statistically downscaled climate projections: is past performance an indicator of future results?

When climate models are developed, researchers test how well they replicate the climate system by using them to model past climate. Ideally, the model output will match the climate conditions that were actually recorded in the past, indicating that the model correctly characterizes how the climate system works and can be used to reliably project future conditions. However, this approach assumes that models that reliably project past climate conditions will accurately predict future climate conditions, even though the climate system might have changed.
 
This research contributes to generating more reliable local-scale climate projections by testing the assumption that the climatological relationships which existed in the past will continue to exist in the future. To do this, researchers developed a novel approach in which very high-resolution climate model data were used as a surrogate for historical and future “observations”, allowing researchers to test how well the more commonly-used coarse-scale global climate models project future climate conditions.

Findings suggest that the assumption holds reasonably well in many cases, but there are some instances (for example in particular geographic locations, such as coastal regions, and at certain times of year, especially summer) when the assumption is not as robust. This research also explores the conditions under which the assumption does not hold, and develops ways to make the methods used to generate local information about climate change more reliable. The results of this research can improve the reliability of the climate models used by resource managers to inform vulnerability assessments, adaptation planning, and other important climate-related decisions.

Comparing and Evaluating Different Models to Simulate Current and Future Temperature and Precipitation

Funding Amount and Duration:

$50,000 from August 1, 2012 - August 31, 2013

Funding Source:

US Geological Survey

Principal Investigators:

Katharine Hayhoe (Co-PI), Texas Tech University

Keith Dixon (Co-PI) & John Lanzante (Co-PI), NOAA Geophysical Fluid Dynamics Laboratory

About:

Regional assessments of the impacts of climate change on both human systems and the natural environment require high-resolution projections to see the effects of global-scale change on the local environment. This project will address a critical and generally overlooked assumption inherent to these projections of regional, multi-decadal climate change: that the statistical relationship between global climate model simulation outputs and real, observed climate data remain constant over time. Utilizing a “perfect—‐model” experimental design and the output of two high-resolution global climate model simulations, this study will evaluate and report on the ability of three different methods to simulate current and future temperature and precipitation in the U.S., with a focus on the southern Great Plains region. Differences between the methods’ abilities during the late 20th versus late 21st century time periods will provide valuable information regarding the level of confidence we should attribute to the climate projections commonly used in impacts analyses and as the basis for decision-support and planning purposes.