Notes from Climate Drought, Snow, and Water I: Room 304-6
Wednesday, May 20, 2015—2pm
Notes by Ben Livneh
I. Robert Oglesby (University of Nebraska, Lincoln) Climate Drivers for Droughts and Pluvials Across the Central US during the Future, Historic, and Pre-Historical Periods: How do human activities affect them?
Three studies presented:
(i) Historic changes in land cover over the Great Plains;
Uses WRF model for two domains, larger (12km), and smaller (4km) uses land cover data sets:
- Current land use
- Pre-settlement vegetation
- Pre-dustbowl vegetation
- Dustbowl vegetation
Land use has changed dramatically. Experiment over Great Plains domain: model driven with the same general meteorology, shows differences in temperature and precipitation associated with changes in land-cover. Further: where it’s warmer there’s less precipitation, these are linked by ET.
(ii) Climate drivers of extended drought and pluvial—comparing SST effects with local effects (soil moisture);
Quantifying the relative roles of local/remote effects on summertime drought. Three different sets of WRF runs with different SSTs were used: (i) positive SST anomalies: conducive to drought; (ii) negative SSTs—conducive to pluvial; (iii) climatological SSTs—neutral. All simulations were initialized in April with low soil moisture.
Result: SSTs have strong control on soil moisture. Low soil moisture from local sources can lead to high pressure aloft and reduced precipitation.
(iii) Improving model simulations of ET: large-scale ET data set for CONUS—in order to model ET, gridded data sets are needed.
Is reduced ET triggering drought, or is drought reducing ET?
METRIC-ET product and WRF have very large disagreement, where irrigation (viewed from satellite and not model) could be a big difference, could be WRF assumed land cover as well.
II. Ruben Behnke (University of Montana) Development of High Resolution Gridded Dew Point Data from Regional Networks
Project goal is to use station data to estimate dew point, rather than infer from day length, temperature, and precipitation as it’s currently done.
History of measuring humidity: there has been a great lack of measurement, hence humidity has been estimated, but recently there has been a huge increase in station abundance since year 1995 it went from < 2000 stations, to more than 16,000 currently.
MT-CLIM is the algorithm developed at U. Montana to estimate humidity empirically and iteratively. It uses solar radiation, topography, aspect, etc... It uses an estimate of aridity to assign dew point: if it’s humid then dew point is set to daily minimum temperature (Tmin); if it’s arid (PET/P>2.5) MTCLIM iterates between humidity and radiation following Kimball et al.
The biases that occur because of this are seasonal and dependent on precipitation, topography, proximity to ocean. Overall predictors vary smoothly, so that diurnal variations tend to create spikes in the bias. At arid locations, the assumption that dewpoint is constant throughout the day breaks down.
To fix this, he recommends switching to using station data for humidity. Great disparity in observations across states, i.e. Ohio for example has stations along its highways.
MADIS is a good forecast product but not great for research. Ruben is trying to collate, QC, and organize the data. Spikes, outliers, station drift, and more make this problematic.
Overall the questions are:
· Is MTCLIM still needed?
· How far back can we go with station data exclusively?
· Is it possible to provide a near-realtime humidity product?
III. Katherine Chase (WY-MT Water Science Center, USGS) Simulated Effects of Climate Change on the Hydrology and Fish Assemblages of Northern Great Plains Streams
Streams are critical places for habitat in this arid region WY-MT. There is a great diversity of fish species and although the sport fish get more attention, the native species are perhaps more interesting. There is a mix of native and “introduced” species. Remarkably, the fish have evolved to survive the wide variety of extreme conditions, drying, cooling, living in small pools or large ones (flooding).
There is a 1600 site data base in Montana. The goal is to examine the role of climate and hydrology on fish using this data base. There is a large gradient in precipitation in Montana. Streamflow was simulated with PRMS for baseline and climate change conditions. Limited calibration was done due to data scarcity, sometimes only seasonal gages were available.
PRMS required precipitation, tmax and tmin from Daymet. Also used NHD for the geospatial fabric. Future climate was driven by RegCM3 with A2 scenario, where a “delta” method was used to difference future temperatures and scale future precipitation from historical.
Interestingly, almost all scenarios showed increased precipitation, while all showed warming temperatures.
Results: The change between current and future streamlow gets wetter or drier depending on which GCM is chosen, i.e. there is a lot of spread, for the first few decades. However, when we get out to 2080, all models agree that flows would get smaller—temperature overwhelms precipitation.
*Applying historical relationships between streamflow and fish biodiversity to future climate is the ultimate goal.
IV. Trace Martyn (University of Wyoming) Past Trends in Snow Accumulation in the Rocky Mountain US and Implications of Future Decreasing Snow Accumulation on the Big Sagebrush (Artemisia Tridentata) Distribution
Looking at past trends in mean minimum temperature (Tmin) and snowdepth on sagebrush plant communities. IPCC report suggests a 1 – 5 degree increase in temperature, while precipitation is much more variable +/- 10%.
IPCC defines Snow-related changes as those that can be detected and attributed to climate change. Snow extent has decreased in the northern hemisphere. It is projected that by 2100 an additional 5-30% decrease in snow extent. She wishes to extend this to snow depth using SNOTEL (automated, daily, since 1970s) and snow course (manual, set courses, weekly, less frequent, since 1920s) data.
Over 1000 stations were used. Minimum temperature only came from SNOTEL, values averaged by year, season.
Results: Showed Tmin regression trends by year by state. To account for unexplained variability, new variables were added: elevation and latitude. Oddly for northern states, Tmin gets warmer with increasing latitude. Mean snow depth appears to be going down for all states, with greater variability in spring versus winter.
Overall, Tmin has increased 0.7 to 2.4 degrees C in the past, which if projected to the future, could resut in an additional 2.6 to 9 degrees C warming. Snowdepth has historically decreased by 4 – 27 cm, which could project out to a decrease 3 – 31 cm in depth for the future.
What does it mean for sage brush? Sage brush requires recharge of the soil, so decreases in snow could reduce the suitability of sage brush to its current habitat.
Discussion:
Management people: What type of questions do you have about climate information and connecting it to what you do?
· Which scales, models, data sets, and time frame are appropriate to make forecasts of ecosystem management and habitat?
o Models likely won’t be able to predict with accuracy climate variables important for fish, but they may be able to project these as a function of GHGs.
· What is needed to make the research more useful and have greater impact; can NCCSC help?
o Input is welcomed, what format is best, what variables are most needed?