Comparing NARR, ERA40, and Observed Hudson Bay River Discharge

1
Comparing NARR, ERA-40, and Observed Hudson Bay
River Discharge Stephen J. Déry1, Eric F. Wood2,
and Christopher Kerr3 1Atmospheric and Oceanic
Sciences Program, Princeton University,
Princeton, NJ 2Department of Civil and
Environmental Engineering, Princeton University,
Princeton, NJ 3Geophysical Fluid Dynamics
Laboratory, National Oceanic and Atmospheric
Administration, Princeton, NJ
Introduction Rivers provide a natural pathway for
freshwater on the land surface and constitute a
vital link between the atmosphere, the land
surface, and the oceans. Streams and rivers
integrate spatially and temporally atmospheric
and land surface processes at the
catchment-scale, providing a mechanism by which
climate change may be detected. One area where
significant climate change is ongoing is the
Arctic (Serreze et al. 2000). Rising surface air
temperatures are altering the hydrologic cycle of
pan-Arctic drainage basins, including freshwater
discharge (Peterson et al. 2002 Déry et al.
2005).One major pan-Arctic river basin that has
undergone recent changes is the Hudson Bay Basin
(which includes the James Bay and Ungava Bay
basins, hereafter referred to as HBB). The HBB
covers an area of 3.7 106 km2, or more than
one third of Canada (Fig. 1). Its freshwater
discharge of 900 km3 yr-1 equates one fifth of
the total annual runoff to the Arctic Ocean and
affects high-latitude oceanographic, atmospheric,
cryospheric, and biologic processes (Aagaard and
Carmack 1989).Given the importance of the HBB
drainage network in the global freshwater budget
and the recent decline in the number of
monitoring flow gauges in Canada (Shiklomanov et
al. 2002), there is an urgent need to report
historical discharge rates of Canadian rivers
with outlets into Hudson Bay. It is also
critical to validate comprehensive
hydrometeorological datasets such as the North
American Regional Reanalysis (NARR) and the
European Centre for Medium-Range Weather
Forecasts (ECMWF) Reanalysis data (ERA-40) in the
polar environment. The goal of this preliminary
study is to compare HBB river discharge
measurements with the modeled streamflow in the
NARR and the ERA-40.
Results (cont.)
Figure 3 presents the annual cycle of daily river
discharge into HBB. The measured flow rates are
relatively low during winter and early spring,
achieving a mean minimum of 0.21 mm day-1 on 5
April. As spring advances, snowmelt accelerates
the rate of discharge and the mean annual maximum
flow rate of 1.37 mm day-1,as a whole, is
reached on average each 25 June. During summer
and early fall, river discharge remains
relatively high, with a secondary plateau
attained on average in late September and early
October. Observed discharge rates then gradually
decrease to the low flow regime of the winter
season.
Figure 3 Mean annual cycle of observed, ERA-40,
and NARR daily river discharge in HBB, 1982-1987.
Compared to the observed state, the reanalysis
products exhibit an amplified seasonal cycle,
with higher daily discharge rates during the
snowmelt period (Fig. 3). The modeled freshets
occur nearly two months prior to the observed
spring peaks. The NARR discharge rates decrease
during the summer and approach zero during
winter. Discrepancies in the annual cycle of
daily discharge arise primarily owing to the lack
of river routing schemes in the reanalysis
products.
Figure 4 illustrates the mean annual observed,
ERA-40, and NARR river discharge per contributing
area in the HBB over the period 1982-1987. The
measured discharge records show higher (lower)
discharge rates per contributing area in eastern
(western) Canada where precipitation rates are
relatively high (low) and evaporation rates are
relatively low (high). Overall, the ERA-40 and
NARR discharge rates exhibit similar spatial
patterns to the observations. For instance,
discharge rates are greater (lesser) on the
eastern shores of Hudson Bay. The NARR captures
the observed streamflow minimum in the Canadian
Prairies and the observed streamflow maximum in
central Québec. The NARR grid point discharge
data tend to be more intense than observed or in
the ERA-40 dataset owing to its finescale
resolution.
Figure 1 River basins of Canada (Source Atlas
of Canada). The Hudson Bay Basin is highlighted
in yellow/orange and covers 3.7 106 km2 in
central Canada.
Data
Figure 4 Mean annual observed (top), ERA-40
(middle), and NARR (bottom) river discharge (mm)
in HBB (bold outline), 1982-1987.
1) OBSERVATIONS Measured freshwater discharge
rates for rivers that drain into HBB are compiled
in Environment Canada's Hydrometric Database
(HYDAT http//www.msc.ec.gc.ca/wsc/hydat/H2O/
index_e.cfm, 2004). HYDAT is a comprehensive
observational database that provides daily
discharge rates for 42 of the main rivers with
outlets into Hudson Bay (see Fig. 1). 2)
ERA-40 The ERA-40 dataset provides instantaneous
discharge rates over the global land surface
generated by a numerical weather prediction model
using a constant analysis framework (ECMWF
http//data.ecmwf.int/data/d/era40_daily/, 2004).
The ERA-40 data used in this study are 6-hourly
values of river runoff on a 1.1o reduced Gaussian
grid. A subset of 266 model grid cells covering
3.5 106 km2 represents the HBB.3) NARR
Instantaneous discharge rates from the NARR are
also used in the analysis (NCEP
http//wwwt.emc.ncep.noaa.gov/mmb/rreanl/, 2004).
The NARR provides 3-hourly surface and
subsurface runoff as generated by the Eta
numerical weather prediction model using a fixed
analysis scheme. A subset of 3400 grid points
from the original NARR mesh (32-km horizontal
resolution) are used to represent HBB river
discharge. The surface and subsurface components
are summed to provide the 1982-1987 daily and
annual HBB river discharge rates that are
compared to the other two datasets.
Conclusion and Future Work
A comparison of the observed, ERA-40, and NARR
river discharge rates in HBB has been
conducted.Our preliminary results show that, for
the period 1982-1987, the NARR annual discharge
rates compare favorably with the observed data
with a mean absolute difference of 31 mm yr-1.
The presented annual cycle of daily discharge
rates from the two reanalysis products exhibits
an amplified seasonal cycle compared to the
observations owing to the lack of a river routing
for the ERA-40 and NARR grid-level runoff. The
NARR discharge rates per contributing area show
similar patterns to the observations, with high
discharge rates occurring in central Québec and
low streamflow rates in the Canadian
Prairies.Further analysis of the NARR
terrestrial hydrologic cycle, using the complete
period (1979-2003), is being undertaken by the
authors and should allow a more complete
assessment of the NARR dataset for hydrologic
studies. Further investigations of large-scale
atmospheric teleconnections and their role in the
HBB hydrologic cycle such as the recent study by
Déry and Wood (2004) will also be conducted.
Results
References Aagaard, K., and E. C. Carmack, 1989
The role of sea ice and other freshwater in the
Arctic circulation. J. Geophys. Res., 94(C10),
14,485-14,498.Déry, S. J., and E. F. Wood, 2004
Teleconnection between the Arctic Oscillation and
Hudson Bay river discharge. Geophys. Res. Lett.,
31, L18205, doi 10.1029/2004GL020729.Déry, S.
J., M. Stieglitz, E. C. McKenna, and E. F. Wood,
2005 Characteristics and trends of river
discharge into Hudson, James, and Ungava Bays,
1964-2000. Submitted to J. Climate.Peterson, B.
J., R. M. Holmes, J. W. McClelland, C. J.
Vörösmarty, R. B. Lammers, A. I. Shiklomanov, I.
A. Shiklomanov, and S. Rahmstorf, 2002
Increasing river discharge to the Arctic Ocean.
Science, 298, 2171-2173.Serreze, M. C., J. E.
Walsh, F. S. Chapin III, T. Osterkamp, M.
Duyrgerov, V. Romanovsky, W. C. Oechel, J.
Morison, T. Zhang, and R. G. Barry, 2000
Observational evidence of recent change in the
northern high-latitude environment. Climatic
Change, 46, 159-207.Shiklomanov, A. I., R. B.
Lammers, and C. J. Vörösmarty, 2002 Widespread
decline in hydrological monitoring threatens
Pan-Arctic research. Eos, Trans. Amer. Geophys.
Union, 83(2), 13.
Figure 2 shows that the mean annual discharge
rate observed in 42 HBB rivers is 235 mm yr-1.
The mean annual discharge rates from the ERA-40
and NARR are 302 and 204 mm yr-1, respectively
(Table 1). The mean absolute errors between for
the ERA-40 and NARR compared to the observations
are 67 and 31 mm yr-1, with the NARR showing a
larger standard deviation than the ERA-40.
Table 1 Statistics on the comparisons between
the annual observed and the ERA-40 and NARR river
discharge rates in HBB, 1982-1987. All units are
mm yr-1. SD, standard deviation MAE, mean
absolute error.
Acknowledgments This research is supported by NSF
grant OPP02-30211 and NOAA grant NA17RJ2612. The
authors thank Ken Mitchell (NCEP), Marco Carrera
(NCEP), and Kirsten Findell (GFDL) for technical
support and their useful comments.
Figure 2 Annual observed, ERA-40, and NARR
river discharge in HBB, 1982-1987.