H53F-01
A Synoptic View of Low Frequency Variability in Fall Precipitation Across the United States
Several recent studies have detected large increases in precipitation and stream flow across the United States over the 20th century, with the largest increases generally being reported in fall precipitation and low flow. Our recent study demonstrated that fall precipitation exhibits strong spatially coherent decadal variations across most of the interior of the country that can be explained by decadal variations in the frequency of rain days. In this study, we examine changes in the fall atmospheric circulation over the United States and identify circulation types that can explain the large trends and decadal variations in the frequency of fall precipitation. We apply a classification and regression trees based approach to categorize the atmosphere into different synoptic patterns. The method uses upper atmospheric and surface variables to classify each fall day into one of several circulation types. This air mass based approach offers several advantages over k means clustering on dynamic variables alone. Most importantly for the current study, the atmospheric patterns have been identified independent of precipitation and may therefore represent synoptic rather than local variations. After classifying each fall day from 1948 to 2004 into one of several synoptic patterns, we disaggregate daily precipitation in different regions of the United States based on circulation type and identify the dominant atmospheric controls on trends and decadal variations in precipitation frequency and total. Our results suggest that the decadal variations in precipitation frequency can be explained by the number of days with a strong trough over the southwestern United States and ridges over the Gulf of Alaska and Great Lakes. This pattern enhances southerly moisture transport from the Gulf of Mexico and the frequency of precipitation in the central United States and Great Plains. Another pattern features a ridge over the southwest and trough over eastern Canada. On these days, meridional flow into the central United States is northerly and the frequency of precipitation is reduced by 80-90%. Changes in the distribution of daily precipitation during different synoptic conditions will be discussed.
H53F-02
The Distribution of Evaporation and Precipitaion
Since precipitation amounts may not be proportional to the evaporation at each location, understanding the distribution of precipitation and evaporation is very important for climate water cycle studies. This study uses data from the version 2 of Goddard Satellite-Based Surface Turbulent Fluxes (GSSTF) data for ocean evaporation. GSSFTF are based on all available SSM/I's (F-8, -10, -11, -13, -14) data. Because there has been no global land energy budget, forcing, and water budget data sets available, four assimilated results (Noah, CLM2, Mosaic, and VIC) from GLDAS (Global Land Data Assimilation System) are analyzed. The rainfall data for both ocean and land are from the Global Precipitation Climatology Project (GPCP). Evaporation over oceans is high over 20- 30 degree regions and lower over equator because of the combination of moisture and wind stress. The evaporation over land are the highest along the equator and dramatically reduced for regions higher than 30 degree.The rainfall are mostly cocentrate and violent along ITCZ and SPCZ, and much lower at regions higher than 25 degree latitudes over both ocean and land.These off-phase distributions between evaporation and precipitation are basically caused by atmosphere dynamics. As expected, there are more evaporation than rain over ocean and the other way around over land. The estimates of precipitation and evaporation over oceans have high accuracy from the available data sets. They could be used to estimate the moisture transports from ocean to land and river runoffs and to validate the land surface models.
H53F-03
Long-term Lake Evaporation Measurements in Southeastern Brazil
We report here for the first time the results of a long-term (37 months) campaign of lake evaporation measurements with the eddy-covariance (EC) method. The measurements were made at Furnas Lake, a large lake (1440 km2) in Southeastern Brazil (20° 44'S, 45° 58'W and 771.8 m ASL). Mean and maximum depths at the Maximum Normal Operating Level are 13 m and 90 m respectively. Taking advantage of a long drought during 2000--2001, a large metal tower was erected over the lake's dry bed. After the water level recovered, we were left with a stable platform for performing EC measurements in one of the lake's many basins. Fetch conditions over the prevailing wind directions were excellent (1000 m from the North, and more than 3000 m from the East), with the closest land at 420 m (from NE) and 440 m (from SW). Measurements included hourly means of water surface temperature, air temperature, specific humidity, downwelling solar radiation, net radiation, wind speed, and wind direction. 10-Hz eddy covariance measurements were made of turbulent fluctuations of 3 wind components, sonic virtual temperature, air temperature (with a fine-wire thermocouple) and of fluctuating specific humidity with a specially adapted capacitive hygrometer. The validation of this sensor to measure latent heat fluxes at high frequency was made on intensive field campaigns that deployed state-of-the art Ultra-Violet and Infra-Red fast-response hygrometers. Our data analysis indicates that atmospheric stability can be far from neutral, and that it plays a very important role in the mass-transfer and heat-transfer equations for the water vapor and sensible heat fluxes. We have also found that significantly different scalar roughenesses for water vapor and for sensible heat were necessary to calibrate properly the Monin-Obukhov Similarity Theory (MOST)-based transfer equations. Due to these differences, gradient-based Bowen ratios (as usually applied in the Energy Budget Bowen Ratio method in the absence of turbulence measurements) do not agree with flux-based Bowen ratios given directly by the ratio of the sensible heat flux and the latent heat flux. Finally, we give the mean monthly values for these two fluxes from July, 2003 to June, 2006 (with 5 months of missing data).
H53F-04
A network of scintillometers for ground-truthing of surface fluxes in New Mexico
A network of seven scintillometer transects was established in New Mexico in 2006 covering different soils, vegetation types, and altitudes. Scintillometers measure spatially-averaged sensible heat fluxes over transects of 0.5 - 4 kms, i.e., over footprints comparable in size to several pixels of a satellite image. The Surface Energy Balance for Land (SEBAL) algorithm is applied to radiances from Landsat and MODIS to obtain net radiation Rnet, soil heat flux G, and sensible heat flux H. The latent heat flux is obtained from the energy balance equation as LE = Rnet - G - H. The scintillometer measurements are used to validate and calibrate the SEBAL sensible heat flux product. Results from this ground-truthing experiment and plans for automated and calibrated daily evapotranspiration maps will be presented.
H53F-05
Analysis of Short-Term Closure of the Surface Energy Balance in Different Seasons
A correct determination of the surface energy balance is an important quality test for measurements of turbulent surface fluxes. The energy balance is often not closed especially in non-homogeneous terrain or in presence of orographic obstacles. Daily energy budget is more easily balanced, because of the contribution of energy residuals of opposite sign; however short-term closure is rarely obtained. What distinguishes this study from previous ones is the effort to close short-term energy budget, and to explore the factors that mainly affect the energy imbalance during the day. To this aim we analysed data sets from southern Italy collected above a semiarid terrain during summer and fall seasons. Our analysis has shown that the global closure rate significantly improves after the correction for the error dependent on the ultrasonic anemometer angle of attack and for the error dependent on the heat storage into the soil. Furthermore a significant reduction of short-term energy residual results by taking into account the contribution to the transport by ‘large scale motions'. The obtained results are independent from the net incoming radiation.
H53F-06
The Role of Irrigation in North American Hydroclimates
Irrigation accounts for nearly 80 percent of the world's consumptive use of fresh water, and irrigated croplands have been shown to influence the terrestrial water balance and land-atmosphere interactions on local to regional scales. However, these hydrologic and climatic effects are not well quantified, particularly at regional to continental scales. Furthermore, accurate initialization of land surface moisture and energy states in numerical weather prediction models is known to enhance short term to seasonal forecast skill, yet there is still room for improvement. To test and quantify the effects of irrigation and explicit crop types on regional to continental scale water and energy budgets over the North American Continent, we recently performed experimental simulations using three land surface models (LSMs), and high-resolution satellite observations of irrigation and crop types. The three LSMs, Noah, Catchment, and the Common Land Model (CLM), are the land components of the operational forecast systems of NOAA, NASA's Global Modeling and Assimilation Office (GMAO), and the National Center for Atmospheric Research, respectively. All three are embedded in NASA's Land Information System (LIS), which facilitates irrigation-related enhancements while reducing redundancy. LIS enables us to drive the LSMs at high resolutions in uncoupled and, later, coupled modes, using observation based parameter and forcing inputs archived by the North American and Global Land Data Assimilation System (LDAS) projects. Preliminary results indicate enhanced evapotranspiration and reduced heating of the land surface over irrigated sites. These enhancements lead to improved predictions of state variables such as surface temperature and soil moisture. Since realistic initialization of land surface states has been shown to be important in weather forecasts, our study provides insight towards the improvement of operational weather and climate predictions through the treatment of irrigation in the land surface components of operational prediction systems.
H53F-07
Do improved hydrological representations in the Community Noah Land Surface Model yield a more robust model?
Understanding the strength of land-memory mechanisms such as the storage of water near the surface as soil moisture and the nature and seasonal progression of growing vegetation remains a challenge. New parameterization schemes aim to capture such processes with realism within land-surface models. We investigate how the augmentation of the latest version of the unified Noah LSM, jointly maintained at NCEP and NCAR, with additional land memory processes (i.e. groundwater and dynamic vegetation) impacts the performance and robustness of the model. It is hypothesized that increased physical realism in conceptual models enhances their robustness, making them less sensitive to the choice of parameter values. At different locations, we systematically perform offline, multiobjective parameter estimation on four different versions of Noah: 1) Standard Noah LSM. 2) Noah-GW, equipped with a simple groundwater scheme, which describes the groundwater dynamics in an unconfined aquifer, efficiently representing the groundwater impacts on soil moisture. 3) Noah-DV, equipped with a short-term dynamic vegetation scheme, which describes the vegetation carbon budgets controlled by photosynthesis and respiration processes and allocating assimilated carbon to roots, stems, wood, and leaves. 4) Noah-DVGW, equipped both with groundwater and dynamic phenology module. The models are constrained with observations of surface energy budget components (sensible, latent, and soil heat fluxes) and soil temperature, moisture, and matric potential. Data was collected from May-June 2002 during the International H2O Project (IHOP-02) at locations representing the major types of land cover in the Oklahoma Panhandle: grassland, pasture, and croplands. The comparison of obtained optimal model structures (e.g., parameter sets, model covariances) allows us to draw conclusions regarding the robustness of the new parameterization to the choice of parameter values. We successively test if new, more realistic model structures significantly change the location of the optimal parameter set with respect to the standard formulation. Such an experimental setup enables us to determine whether addition of new parameters alters the structure of the model such that parameter values for existing parameterizations change. Additionally we can single out the contribution of individual parameterization enhancements to the overall improvement in performance.
H53F-08
Short to Medium-Range Hydrometeorological Forecasts in the Rio Grijalva Basin, Mexico
This paper describes research in support of a project to enable the interpretation of modeled meteorological fields in terms of streamflow in the Rio Grijalva basin, located in southern Mexico. The Rio Grijalva basin is the most important basin in terms of hydropower production, and one of the basins most affected by floods in Mexico. So establishing a short to medium-range hydrometeorological forecasting system is recommended. A physical, distributed, hydrological model (MMS-PRMS) is established through the following steps: 1) basin parameterization, 2) parameter optimization, and 3) implementation of modeled meteorological fields into the resulting hydrological model. Most datasets for topographic, soil and vegetation parameter derivation for the MMS- PRMS are only available in the United States so an alternative derivation methodology from global, publicly available, surrogate datasets is proposed. Parameter optimization is performed through the Shuffled Complex Evolution method with the use of a local hydrometeorological network. The documentation of these initial steps is considered relevant for other hydrological modelers in Mexico and other countries where hydrological models, parameterization datasets, and optimization tools are limited. The short-term predictive capabilities of the resulting model are tested using modeled rainfall and temperature from the North American Regional Reanalysis (NARR). A relevant bias in NARR-rainfall is identified. Methodologies for a probabilistic bias-correction and uncertainty estimation in the meteorological fields are proposed. The bias-identification and correction are perhaps the most important results. Thus suggesting NARR fields should be should follow a similar process previous to their analysis.