CAHDMA3 Extended Abstract Template

Document Sample
CAHDMA3 Extended Abstract Template Powered By Docstoc
					  Historical Simulations of Land Surface Conditions and Their
  Applications to Analyses of Decadal Hydrological Changes
                          Aiguo Dai, Taotao Qian, and Kevin E. Trenberth

                National Center for Atmospheric Research, Boulder, Colorado, USA

Abstract Historical data for soil moisture, evaporation, and other surface fields are sparse or
unavailable over most land areas, yet they are needed for many climate change analyses. Here we
demonstrate that simulations using land surface models forced by observed precipitation,
temperature, and other atmospheric forcing can supplement available data for analyses of changes
in the hydrological cycle on regional and continental scales.

Introduction, Forcing Data and Model
     Historical records of surface evaporation, runoff, soil moisture, and other land surface fields are
unavailable over most of the continents. Because of this, simulations of historical land surface
conditions using land surface models are needed for studying variability and changes in the
continental water cycle and for providing initial conditions for seasonal climate predictions. Besides
the fidelity of the models, realistic atmospheric forcing data are crucial for these simulations,
especially for climate change applications (Ngo-Duc et al. 2005; Qian et al. 2006; Sheffield et al.
2006). For this purpose, we have constructed a global forcing data set for 1948-2004 with 3-hourly
and T62 (~1.875o) resolution, made global simulations for 1948-2004 using the Community Land
Model version 3.0 (CLM3), evaluated the simulations using available observations of streamflow,
continental freshwater discharge, surface runoff, and soil moisture, and applied the simulated data
in hydroclimatic trend analyses for the Mississippi River basin (Qian et al. 2006; Qian et al. 2007).
     The forcing data set was derived by combining observation-based analyses of monthly
precipitation and surface air temperature with intra-monthly variations from the National Center for
Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) reanalysis,
which is shown to have spurious trends and biases in surface temperature and precipitation. Surface
downward solar radiation from the reanalysis was first adjusted for variations and trends using
monthly station records of cloud cover anomaly and then for mean biases using satellite
observations during recent decades. Surface specific humidity from the reanalysis was adjusted
using the adjusted surface air temperature and reanalysis relative humidity. Surface wind speed and
air pressure were interpolated directly from the 6-hourly reanalysis data. Sensitivity experiments
show that the precipitation adjustment (to the reanalysis data) leads to the largest improvement,
while the temperature and radiation adjustments have only small effects.
     The CLM is a comprehensive land surface model designed for use in coupled climate system
models. It is described in detail by Oleson et al. (2004). In the CLM, spatial heterogeneity of land
surface is represented as a nested sub-grid hierarchy in which grid cells are composed of multiple
land units, snow/soil columns, and plant functional types (PFTs). Each grid cell can have a different
number of land units, each land unit can have a different number of columns, and each column can
have multiple PFTs. Biogeophysical processes are simulated for each sub-grid land unit, column,
and PFT independently and each sub-grid land unit maintains its own prognostic variables. The
grid-averaged atmospheric forcing is used to force all sub-grid units within a grid cell.
     When forced by the observational data set, the CLM reproduces many aspects of the long-term
mean, annual cycle, interannual and decadal variations, and trends of streamflow for many large
rivers (e.g., the Orinoco, Changjiang, Mississippi, etc.), although substantial biases exist (Fig. 1).
The simulated long-term mean freshwater discharge into the global and individual oceans is
comparable to 921-river-based observational estimates. Observed soil moisture variations over
Illinois and parts of Eurasia are generally simulated well, with the dominant influence coming from
precipitation (Qian et al. 2006). The results suggest that the CLM simulations are useful for climate
change analysis.

FIG.1. CLM-simulated (dashed) and the observed (solid) water-year (Oct.-Sep.) river outflow rates for world’s 10
largest rivers (except No.5 Brahmaputra and No. 10 Mekong whose records are too short), and four smaller rivers that
have long records. Also shown on top of each panel are long-term means (from left to right) of the observed and CLM-
simulated rates, and the correlation coefficient (r) between the two curves.

Hydroclimatic Changes Over the Mississippi River Basin

     Figure 2 shows the basin-averaged time series from 1949-2004 of annual (water year: Oct-Sep)
precipitation, evaporation, runoff, and land water storage (W), together with the surface heat fluxes
(surface shortwave radiation SW, longwave radiation LW, latent heat flux LH, and sensible heat
flux SH). For the surface water budget, the observed increase in basin-averaged precipitation is
compensated by increases in both runoff and evapotranspiration. For the surface energy budget, the
decrease of net shortwave radiation associated with observed increases in cloudiness is
compensated by decreases in both net longwave radiation and sensible heat flux, while the latent
heat flux increases in association with wetter soil conditions. Both the surface water and energy
budgets support the view that evapotranspiration has increased in the Mississippi River basin from
1948-2004. Sensitivity experiments show that the precipitation and temperature changes dominate
the evapotranspiration trend, while the solar radiation change has only a small effect. Large spatial
variations within the Mississippi River basin and the contiguous United States are also found.
However, the increased evapotranspiration was ubiquitous despite spatial variations in

FIG. 2. Time series of annual (water-year) surface water (left) and energy (right) budget components averaged over the Mississippi
River basin and the associated linear trends (straight lines, b is the slope). Except for the observed precipitation (P, solid line) and
runoff at Vicksburg, Mississippi (Robs, solid line with triangles) and the water budget-derived evapotranspiration (P-Robs-dW/dt, solid
line with dots), other components are from CLM3 simulations. (Adapted from Qian et al. 2007)

Changes in Continental Freshwater Discharge from 1949-2004

     The CLM3-simualted streamflow were used to fill the data gaps in the historical records of
streamflow at the farthest downstream stations for world’s largest 925 rivers using a regression-
based procedure described in Dai et al. (2007). The simulated runoff fields were used to estimate
the discharge from drainage areas not covered by this river network. The station flow rates were
adjusted to represent river mouth outflow rates using the ratio of the simulated flow at the station
and river mouth.
     Figure 3 shows large variations in continental discharge at interannual to decadal time scales.
These variations are correlated with the Southern Oscillation Index (SOI) for the discharge into the
Pacific, Atlantic, Indian, and global oceans as a whole (but not with discharges into the Arctic
Ocean and the Mediterranean and Black Seas), consistent with previous regional analyses.
Consistent with previous reports, there is a large upward trend in the water-year discharge into the
Arctic Ocean (~0.21 × 10-3 Sv yr-1) from 1949-2004. For the other ocean basins and the global
oceans as a whole, the discharge data show downward trends, which are statistically significant for
the Pacific (−0.43 × 10-3 Sv yr-1) and global oceans (−0.58 × 10-3 Sv yr-1). Except for the Arctic
discharge, precipitation changes are found to be the main cause for the discharge trends and large
interannual to decadal variations, although the CLM3 simulation also suggests influences of surface
temperature and other atmospheric forcing (e.g., through enhanced evaporation). For the Arctic
drainage areas, the upward trends in streamflow are not accompanied by increasing precipitation,
especially over the Siberia. The CLM3 simulation suggests that recent surface warming has induced
decreasing trends in snow cover and soil ice water content over the northern high-latitudes, which
contribute to the runoff increases in these regions. Our results contradict the notion that continental
runoff has increased during the recent decades.

FIG.3. Time series of annual (water-year) freshwater discharge (solid line, in Sv, 1 Sv=106 m3 s-1) from land into the
individual and global oceans from 1949-2004. The shading indicates the ± one standard error, which includes the
regression error and the observational error (estimated as the difference between the observed and the estimated river
flow using the regression equation and CLM-simulated flow). Also shown (dashed) is the Southern Oscillation Index
(SOI, following Trenberth 1984) averaged over the 12 month period that yields a maximum correlation (r) with the
discharge data (see text for details). The mean (m), linear slope (b) and its attained significance level (p(b)) of the
discharge time series are given on top of each panel. (from Dai et al. 2007).


Dai, A., T. Qian, and K. E. Trenberth, 2007: Changes in continental freshwater discharge from
     1949-2004. J. Climate, submitted.
Ngo-Duc, T., J. Polcher, and K. Laval, 2005: A 53-year forcing data set for land surface models.
       Journal of Geophysical Research-Atmospheres, 110, D06116, doi:10.1029/2004JD005434.
Qian, T., A. Dai, K. E. Trenberth, and K. W. Oleson, 2006: Simulation of global land surface
     conditions from 1948-2004. Part I: Forcing data and evaluation. J. Hydrometeorology, 7,
Qian, T., A. Dai, and K. E. Trenberth, 2007: Hydroclimatic Trends in the Mississippi River Basin
     from 1948-2004. J. Climate, in press.
Sheffield, J., G. Goteti, and E. F. Wood, 2006: Development of a 50-year high-resolution global
     dataset of meteorological forcings for land surface modeling. Journal of Climate, 19, 3088-