- Program Overview
- Program Goals
- Program Challenges
The Colorado Headwaters effort was initiated in the Spring of 2008 as a project within the Research Applications Laboratory's Water System program. It is focused on assessing the impact of climate change on winter precipitation, snowpack and runoff processes in Colorado's headwater basins using a very high resolution fully coupled atmospheric–hydrologic model (WRF coupled with the NOAH land surface model). This work is collaborative with NCAR/MMM and researchers from the University of Vienna, George Mason University, and the University of Texas.
Figure 1: The WRF model domain and location
of SNOTEL sites (black dots). (a) The full model
domain. (b) Sub–domain focused on the 112
SNOTEL sites in the Colorado Headwaters region.
The project was awarded 500,000 computer GAUs through an Accelerated Science Discovery competition in the spring of 2008. Using these resources, the Advanced Research WRF (ARW) model was applied to conduct high–resolution regional climate simulations of cold–season snowfall, snowpack, evapotranspiration and runoff in the Colorado Headwaters region. The domain of the high–resolution model is shown in Figure 1a, and the SNOwpack TELemetry (SNOTEL) observation sites used to evaluation the simulations in Figure 1b as the black dots. The specific simulations performed were as follows: i) five 6–month (1 November to 1 May), 2–km–resolution simulations of present–day climate using the North American Regional Reanalysis (NARR) data, covering two cold seasons (i.e., year 2004–2005, and year 2005–2006) of normal (i.e., approximately multi–year mean) precipitation and snowpack,
Figure 3: Sub–domain average precipitation
accumulation from the 2–km WRF model simulation
using retrospective NARR data (blue line) and
perturbed NARR data (PGW simulation; red line).
The results are from the 6–month period from
(a) 2001–2002, (b) 2003–2004, (c) 2005–2006,
and (d) 2007–2008. Click to Enlarge one cold season (i.e., year 2007–2008) of anomalously high snowfall and snowpack, one cold season (i.e., year 2002–2003) of anomalously low snowfall and snowpack, and the 2008 warm season; ii) a few coarse–resolution simulations of the 2007–2008 cold season with grid spacings of 6 km, 18 km and 36 km; iii) four 6–month, 2–km–resolution simulation of snowfall and snowpack in response to a pseudo climate warming, in which the initial and boundary conditions were derived from the combination of the 3–hourly NARR data from the four retrospective cold seasons listed in (i) with the climate perturbations representing the differences between the present (i.e., 10–year averages from 1995–2005) climate and the future (i.e., 10–year averages from 2045–2055) climate projected by CCSM; and iv) two future climate simulations at grid spacings of 6km and 18 km for the 2052–2053 cold season using the 6–hourly CCSM output for the IPCC SRES A1B scenario. During the FY2010, Pseudo Global Warming (PGW) simulations were performed for the 2007–2008 cold season on a 36 km in order to study the impact of simulated climate signal on model resolution.
A team of scientist from a variety of disciplines such as atmospheric science, hydrometeorology, climate and regional modeling, land–surface modeling, and social science will collaborate to investigate following questions:
- How high should model resolution be for the regional climate model to accurately simulate seasonal snowfall and snowpack in the Colorado Headwaters regions?
- Will the predicted increase in snowfall (caused by a warmer, moister climate) be enough to offset the accelerated melting and sublimation due to warmer temperatures?
- If so, will the increase in snowfall be sufficient to maintain river flow at current levels?
Results from four six–month high–resolution WRF simulations were compared with the SNOwpack TELemetry (SNOTEL) data. The results showed that the model properly simulated seasonal snowfall. However, the model underestimated snowpack. Onset and amount of melting and sublimation of snowpack depend on several factors such as terrain features, land surface type, and shadowing effect. Thus, snowpack simulation is sensitive to physics and parameterizations used in the land surface model. Our team is working on improving the land–surface model to accurately simulate snowmelt/snowpack.