• Aagaard, K., and E. C. Carmack, 1989: The role of sea ice and other fresh water in the Arctic circulation. J. Geophys. Res.,94, 14 485–14 498.

  • Broecker, W. S., D. M. Peteet, and D. Rind, 1985: Does the ocean–atmosphere system have more than one stable mode of operation? Nature,315, 21–26.

  • Bryazgin, N. N., 1976: Yearly mean precipitation in the Arctic region accounting for measurement errors (in Russian). Proc. Arctic Ant. Res. Inst.,323, 40–74.

  • de la Mare, W. K., 1997: Abrupt mid-twentieth century decline in Antarctic sea ice extent from whaling records. Nature,389, 57–60.

  • Dickson, R., J. Lazier, J. Meincke, P. Rhines, and J. Swift, 1996: Long-term coordinated changes in the convective activity of the North Atlantic. Progress in Oceanography, Vol. 38, Pergamon Press, 241–295.

  • Greenspan, G., 1974: Discrete Numerical Methods in Physics and Engineering. Vol. 107, Mathematics in Science and Engineering, Academic Press, 312 pp.

  • IPCC, 1996: Climate Change 1995: The Science of Climate Change, J. T. Houghton et al., Eds., Intergovernmental Panel on Climate Change, Cambridge University Press, 572 pp.

  • Karl, T., 1998: Regional trends and variations of temperature and precipitation. The Regional Impacts of Climate Change: An Assessment of Vulnerability, R. T. Watson, M. C. Zinyowera, R. H. Moss, and D. J. Dokken, Eds., Intergovernmental Panel on Climate Change, Cambridge University Press, 412–425.

  • Kattsov, V. M., J. E. Walsh, A. Rinke, and K. Dethloff, 2000: Atmospheric climate models: Simulations of the Arctic Ocean fresh water budget components. The Freshwater Budget of the Arctic Ocean, E. L. Lewis, Ed., Kluwer Academic, in press.

  • Khrol, V. P., Ed., 1996: Atlas of Water Balance of the Northern Polar Area. Gidrometeoizdat, 81 pp.

  • Legates, D. R., and C. L. Willmott, 1990: Mean seasonal and spatial variability in gauge-corrected global precipitation. Int. J. Climatol.,10, 111–133.

  • Parker, D. E., and M. Jackson, 1995: The standard GISST data sets:Versions 1 and 2. Workshop on Simulations of the Climate of the Twentieth Century Using GISST, Bracknell, United Kingdom, Hadley Centre, 50–51.

  • Radionov, V. F., N. N. Bryazgin, and E. I. Alexandrov, 1997: The Snow Cover of the Arctic Basin. Applied Physics Laboratory Tech. Rep. APL-UW TR 9701, University of Washington, Seattle, WA, 95 pp. [Available from Applied Physics Laboratory, University of Washington, 1013 NE 40th Street, Seattle, WA 98105.].

  • Roeckner, E., and Coauthors, 1996: The atmospheric general circulation model ECHAM-4: Model description and simulation of present-day climate. Max Planck Institute for Meteorology Rep. 218, 90 pp. [Available from Max-Planck-Institut für Meteorologie, Bundesstr. 55, D-20146, Hamburg, Germany.].

  • Vinje, T., 1999: Barents Sea ice edge variations over the past 400 years. Proc. Workshop on Sea Ice Charts of the Arctic, Seattle, WA, World Climate Research Programme/Arctic Climate System Study, 4–6.

  • WCRP, 1996: Arctic Climate System Study (ACSYS): Initial implementation plan. World Climate Research Programme WCRP-85 WMO/TD-No. 627, 66 pp. [Available from International ACSYS Project Office, Polar Environmental Centre, N-9296 Tromso, Norway.].

  • Zakharov, V. F., 1997: Sea ice in the climate system. World Climate Research Programme/Arctic Climate System Study WMO/TD-No. 782, 80 pp.

  • View in gallery

    Mean seasonal cycles and ranges of interannual variation during the period 1949–89 of area-averaged precipitation (mm day−1) for the ocean area poleward of 70°N simulated by ECHAM-4 (heavy solid line) relative to observational estimates of Bryazgin (1976) (B) and Legates and Willmott (1990) (L).

  • View in gallery

    Time series of annual precipitation for 55°–85°N: (a) ECHAM-4 simulated anomalies relative to 1961–90 model mean, and (b) observationally derived anomalies relative to 1961–90 observational means, as shown by IPCC (1996). Smooth curves are obtained from nine-point binomial filtering of annual values; thin curve in (a) is based on data for land only.

  • View in gallery

    As in Fig. 2 but for Karl’s (1998) “Arctic region”; observational data in (b) are from Karl (1998).

  • View in gallery

    As in Fig. 2b but based on a Laplacian interpolation in areas for which no observational data were available.

  • View in gallery

    Linear (least squares fit) trend of annual precipitation in ECHAM-4 model simulation. Trends are expressed as the percentage change (over 92 yr, 1903–94) relative to the 1961–90 mean at each grid point.

  • View in gallery

    Sea-ice extent anomalies, relative to 1903–94 mean, prescribed in the ECHAM-4 model simulation: (a) Northern Hemisphere and (b) Southern Hemisphere. Nine-point binomial filter has been applied to each time series.

  • View in gallery

    Sea surface temperature anomalies, relative to 1903–94 mean, in the ECHAM-4 model simulation: (a; top) Northern Hemisphere and (b; bottom) Southern Hemisphere. Nine-point binomial filter has been applied to each time series.

  • View in gallery

    Linear (least squares fit) trends of annual sea surface temperature in GISST dataset, 1903–94. Trends are expressed in °C per 92 yr.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 232 232 27
PDF Downloads 91 91 10

Twentieth-Century Trends of Arctic Precipitation from Observational Data and a Climate Model Simulation

View More View Less
  • 1 Voeikov Main Geophysical Observatory, St. Petersburg, Russia
  • | 2 Department of Atmospheric Sciences, University of Illinois, Urbana–Champaign, Urbana, Illinois
© Get Permissions
Full access

Abstract

The ECHAM-4 global climate model, forced by time-dependent ocean boundary conditions and CO2 concentrations, shows a substantial increase of precipitation in the Arctic during a simulation of the twentieth century. Observational data from the Intergovernmental Panel on Climate Change and other sources show a similar increase. The largest increases occur at nearly the same time in the observational data and the model output, implying that the ocean boundary conditions may be the primary driver of the increase. However, attribution to ocean boundary conditions is subject to the caveat that the largest increase of simulated precipitation occurs near the sea-ice margin, which contained essentially no interannual variability in the ECHAM-4 simulation during the first half-century (for which sea-ice data were sparse), after which time the model’s sea-ice extent was prescribed to decrease substantially. Uncertainties in the precipitation data must also be addressed further before the observationally derived trend of precipitation can be considered to be robust.

Corresponding author address: Dr. John E. Walsh, 102 Atmospheric Science Building, University of Illinois, 105 S. Gregory Ave., Urbana, IL 61801.

Email: walsh@atmos.uiuc.edu

Abstract

The ECHAM-4 global climate model, forced by time-dependent ocean boundary conditions and CO2 concentrations, shows a substantial increase of precipitation in the Arctic during a simulation of the twentieth century. Observational data from the Intergovernmental Panel on Climate Change and other sources show a similar increase. The largest increases occur at nearly the same time in the observational data and the model output, implying that the ocean boundary conditions may be the primary driver of the increase. However, attribution to ocean boundary conditions is subject to the caveat that the largest increase of simulated precipitation occurs near the sea-ice margin, which contained essentially no interannual variability in the ECHAM-4 simulation during the first half-century (for which sea-ice data were sparse), after which time the model’s sea-ice extent was prescribed to decrease substantially. Uncertainties in the precipitation data must also be addressed further before the observationally derived trend of precipitation can be considered to be robust.

Corresponding author address: Dr. John E. Walsh, 102 Atmospheric Science Building, University of Illinois, 105 S. Gregory Ave., Urbana, IL 61801.

Email: walsh@atmos.uiuc.edu

1. Introduction

The polar amplification of global warming is the most widely cited feature of the greenhouse climate change projected for the Arctic. Of potentially greater global importance, however, are the hydrologic changes that are likely to accompany an Arctic warming. The subarctic seas are the locations of the deep convection that is a primary driver of the global thermohaline circulation of the ocean (Broecker et al. 1985), and the rates and locations of oceanic convective activity are sensitive to the freshwater budget (and hence the stratification) of the subpolar seas. The regions from which freshwater advects into the subpolar seas are also important in this context. Through the outflow of sea ice and relatively fresh surface-layer water, the Arctic Ocean and its terrestrial watersheds ultimately provide much of the freshwater that reaches the upper layers of the Greenland, Iceland, and Labrador Seas. These seas appear to be“delicately poised with regard to their ability to sustain convection” (Aagaard and Carmack 1989, p. 14 485), and there are indications that their rates of deep convection have indeed varied substantially over recent decades (Dickson et al. 1996). Given the importance of this convection for the deep ocean circulation, it follows that changes in the Arctic freshwater budget can have profound effects on the broader Atlantic and even the global ocean. In this article, we bring together recent model results and observational data compilations to show that they present a generally consistent picture of a twentieth-century increase of Arctic precipitation. However, the uncertainties that accompany the model experiments and the data compilation are sufficiently large that conclusions concerning the increase of Arctic precipitation can only be regarded as preliminary at this time. Yet the potential importance of this increase appears to call for more systematic model experimentation in order to address the validity and origin of any recent decade-to-century-scale changes of Arctic precipitation. In the following sections we present initial results of a model–data comparison, indicate the major limitations of the work to date, and suggest strategies for narrowing the uncertainties in the trends of Arctic precipitation over the past century

2. Model–data comparison

The model results examined here are from a simulation of twentieth-century climate by the ECHAM-4 global atmospheric model (Roeckner et al. 1996). In the simulation, the sea surface temperature and sea-ice coverage were prescribed on a monthly basis in accordance with the historical observations synthesized in the Global Ice and Sea Surface Temperature (GISST) dataset (Parker and Jackson 1995) of the U.K.’s Hadley Centre. (As discussed below, the ECHAM group made some noteworthy modifications to the GISST sea ice). In addition, atmospheric CO2 concentrations in the ECHAM-4 run were prescribed to follow a temporal evolution consistent with that observed during the twentieth century. Thus the effects of the CO2 (greenhouse) and surface-temperature forcings are not separable in the output of this single simulation.

The climatology of the ECHAM-4 simulation replicates the approximate magnitudes and seasonality of the observationally derived estimates of Arctic precipitation by Bryazgin (1976) and Legates and Willmott (1990). Figure 1 shows that the annual cycles of both the ECHAM-4 precipitation and the observational estimates for the polar cap (70°–90°N) contain peaks of 1.2–1.3 mm day−1 in late summer/early autumn. The ECHAM-4 precipitation rates are 10%–35% larger than the observational estimates during most of the winter months, although it must be noted that 1) the problem of gauge-undercatch is most severe during winter, and 2) the observational estimates are within the range of the interannual variations of the simulated amounts during most calendar months. Only in the March–May period, during which the observational estimates are smallest, do the observational estimates consistently fall outside the model’s interannual range. The ECHAM-4 model’s annual mean of 0.78 mm day−1 is almost exactly midway between the annual means of the two observational sources: 0.75 mm day−1 (Legates-Willmott 1990) and 0.80 mm day−1 (Bryazgin 1976).

Figure 2a is a time series of the annual anomalies (relative to the 1961–90 model mean) of areally averaged precipitation for 55°–85°N from the ECHAM-4 model simulation. Figure 2a shows that a temporal trend is superimposed on the interannually variable precipitation amounts. The smoothed (by a nine-point binomial filter) time series shows an increase of approximately 40 mm from the early 1900s to the early 1990s. This increase represents a change of approximately 8% from the simulated annual mean for the early decades of the century. The most rapid increase occurs during the 1920s, although the filtered time series continues to show a net increase until the early 1990s. The dip during the mid-1990s may be related to an artificial specification of ocean boundary conditions for the final years of the simulation (see below).

To permit a more direct comparison with the observational dataset that is based on land-station data, Fig. 2a also shows (thin line) the filtered time series of the model-derived precipitation anomalies averaged over only the land areas within 55°–85°N. The “land-only” curve differs only slightly from the “all-area” curve, implying that the simulated precipitation over the data-sparse ocean area within 55°–85°N has little effect on the overall character of the time series. It must be noted, however, that the Arctic land station network is biased toward coastal locations and low elevations. These stations may not be representative of the broader northern land areas.

Figure 2b shows the corresponding observationally derived time series from the IPCC (1996) for the identical latitude band. The similarity to Fig. 2a is striking. The filtered time series in Fig. 2b shows a slightly larger (50–60 mm) increase than does the model’s filtered time series in Fig. 2a, although both show the largest increase during the decade of the 1920s. Some secondary features such as the dip around 1970 and the peak around 1990 are also common to both time series.

To substantiate the agreement between the two sources of precipitation information, Fig. 3 shows a similar comparison for Karl’s (1998) “Arctic region,” which is the area poleward of 65°N but excluding Iceland and including the waters surrounding southern Greenland (Karl 1998, his Fig. A-3). The primary difference from Fig. 2 is that both time series in Fig. 3 show a steadier increase from 1900 to the 1950s, that is, the abruptness of the increase during the 1920s in Fig. 2 is no longer present. Again, the magnitude of the increase of the simulated values is somewhat smaller than that of the observational time series.

While the agreement between the model- and data-derived time series is impressive at first glance, several important limitations may affect the validity of the comparison. A first significant concern, which pertains to the observational time series, is the adequacy of the station network from which the areal means were computed. For example, while the station network is relatively dense between 55° and 70°N, the number of stations decreases rapidly north of 70°N. In the Global Historical Climate Network (GHCN), which includes nearly all the stations used by the IPCC and Karl (1998), the number of 5° × 5° grid cells containing observations ranges from a maximum of 40–60 during the 1960–90 period to as few as 6–12 in some months during 1900–20. In an attempt to address the issue of data representativeness, we have applied a Laplacian interpolation (e.g., Greenspan 1974, p. 80) to the GHCN data in order to fill in the spatial gaps in the observational data (primarily in the ocean areas, including the Arctic Ocean). The interpolation procedure included a specification of monthly climatological values at the North Pole from N. Bryazgin’s observational compilation (see Khrol 1996); by anchoring the interpolation in this way, we avoid the positive bias that would result if the Arctic Ocean’s precipitation were interpolated solely from surrounding terrestrial values. Figure 4 shows the modified version of the observational time series for 55°–85°N after inclusion of the interpolated values. Two notable differences from Fig. 2b are 1) the negative anomalies of the early decades are somewhat smaller in magnitude, resulting in a smaller net increase during the twentieth century (the increase is now more comparable to the model-derived change), and 2) the largest increase occurs near the middle of the century in Fig. 4 rather than in the 1920s as in Fig. 2b. The differences imply that there is considerable uncertainty in the observational time series; as noted earlier, this uncertainty is likely intertwined with the temporal and spatial distribution of the observing stations. Nevertheless, the filtered time series in Fig. 4 is broadly consistent with Radionov et al.’s (1997, their Fig. 31) independently compiled time series of precipitation over the Arctic Ocean. Both time series show relative maxima in the mid-to-late 1960s, relative minima in the early 1970s, maxima in the late 1970s, and relative minima in the mid-1980s. For the 32-yr period (1957–88) spanned by Radionov et al.’s data, the correlation between the two time series is r = +0.52.

A second major concern arises from the specification of the model’s lower boundary conditions, which affect the spatial pattern of the model’s precipitation trends (Fig. 5). It is apparent from Fig. 5 that the increase of Arctic precipitation is greatest in the marginal sea-ice zone (MIZ), particularly the summer MIZ north of Eurasia and North America. In fact, the only other areas of the globe having stronger trends than the MIZ are the tropical oceans and the Red Sea region. Although the trends over the northern terrestrial regions of Asia and North America are not as strong as those over the MIZ, they are nonetheless positive as indicated by the yellow-green tones in Fig. 5 and by Kattsov et al.’s (2000) Figs. 14b and 14c, which show twentieth-century increases of 20–30 mm in the temporally filtered time series of Asian and North American Arctic terrestrial precipitation from the ECHAM-4 simulation. The concentrated increase of precipitation near the sea-ice margin implies a decrease of sea-ice coverage, which is strikingly apparent in Fig. 6. However, Fig. 6 contains several unrealistic features: the constancy of the sea-ice boundaries prescribed from 1900 until the 1940s (the ECHAM group prescribed a sea-ice climatology in the absence of reliable information on sea-ice variations during this period); the abrupt decrease of sea ice in the 1940s when the more reliable data on sea-ice variability became available; and the rapid increase during the last few years, when the older sea-ice climatology was again employed in the absence of available sea-ice data for the most recent of the simulated years. With regard to the rapid decrease of sea ice near the middle of the simulation, we note that observational data (not used in this experiment) indicate that a large decrease of sea ice may indeed have occurred during the early-to-middle of the century in the North Atlantic (Zakharov 1997; Vinje 1999) and in the Antarctic (de la Mare 1997). However, the underlying datasets are only now becoming available for systematic comparisons with the more recent and robust datasets. It is apparent that further simulations of twentieth century atmospheric variability can and should utilize improved sea-ice data for 1900 through the 1940s and for the 1990s, especially since Fig. 3b (although not Fig. 2b) shows that the observationally derived precipitation increased most rapidly in the 1950s, considerably later than the abrupt decrease of sea ice. It is also noteworthy that Fig. 5 does not show a particularly large increase of precipitation over the North Atlantic MIZ or over the wintertime MIZ in the North Pacific.

Given that the model’s sea ice did not vary interannually until the late 1940s, the only external forcings that could produce an increase of the model’s precipitation during the early decades were the specified SST and CO2. Figure 7a shows that the areally averaged SST of the Northern Hemisphere increased primarily during 1910–60, in a manner consistent with the increase of Arctic precipitation. The variations of SST are similar in all seasons, although the net increase from 1910 to 1960 is slightly larger in summer than in other seasons. The general temporal pattern of SST in Fig. 7a corresponds well with the twentieth century variations of surface air temperature (IPCC 1996, p. 143), for which the century-scale increase was interrupted by a cooling during the 1960s and early 1970s. The Southern Hemisphere time series of SST (Fig. 7b) shows a steadier rate of increase throughout the century, in agreement with the estimates of surface air temperature for the Southern Hemisphere. (The return to the mean values at the end of the time series in Fig. 7 is again due to the specification of climatological values for the final few years of the simulation.)

In both hemispheres, the largest increases of SST are found in the subpolar regions (Fig. 8). Particularly noteworthy here are the relatively large increases near the sea-ice margin and the subpolar oceans immediately to the south. The increases in the former are associated with the reduction of sea-ice cover, especially in the Eurasian and Pacific sectors. However, the increases of SST in the ice-free portions of the subpolar North Pacific and North Atlantic suggest that an increase of evaporation equatorward of the ice margin may also have contributed to the Arctic precipitation enhancement.

3. Conclusions

The combination of the spatial pattern of SST in Fig. 8 and the temporal correspondence between Figs. 6a–7a and Figs. 2, 3 strongly suggest that at least some (and probably most) of the twentieth-century increase of Arctic precipitation is attributable to the variations of the sea surface temperature and sea-ice boundary conditions. However, the increase of simulated (and observed) Arctic precipitation may not be independent of the CO2 increase if the latter has contributed to the general warming of the subpolar oceans and the marginal sea-ice zone. Model experiments in which the forcing effects of SST, sea ice, and CO2 are isolated in individual simulations are required to evaluate the direct contributions of the lower boundary conditions and the direct radiative effects of the CO2 increase. Such experiments are now feasible, and one intention of this paper is to highlight the need for such experiments in the context of the freshwater budget of the Arctic and the subpolar seas.

The results presented here also show the need for greater attention to 1) the specification of sea-ice boundary conditions in model experiments and 2) the estimation of area-averaged precipitation for the Arctic. Recent digitizations of sea-ice data for the early twentieth century imply that substantial improvements with regard to 1) may now be possible. Progress with regard to 2) may also be imminent through the hydrologic subprogram of the Arctic Climate System Study (WCRP 1996), although the temporal heterogeneity of the historical data on high-latitude precipitation and related variables poses significant challenges to the construction of more robust time series of twentieth-century Arctic precipitation.

Acknowledgments

The MGO portion of this study was supported by the Russian Foundation for Basic Research through Grant 96-05-65020. The Illinois portion was supported by the National Science Foundation through Grants ATM-9319952 and ATM-9612324. Klaus Arpe of the Max Planck Institute and Petr Sporyshev of the Main Geophysical Observatory made available the ECHAM-4 model output used in this study. The GHCN dataset was provided by Tom Karl of NCDC with the assistance of Dick Knight.

REFERENCES

  • Aagaard, K., and E. C. Carmack, 1989: The role of sea ice and other fresh water in the Arctic circulation. J. Geophys. Res.,94, 14 485–14 498.

  • Broecker, W. S., D. M. Peteet, and D. Rind, 1985: Does the ocean–atmosphere system have more than one stable mode of operation? Nature,315, 21–26.

  • Bryazgin, N. N., 1976: Yearly mean precipitation in the Arctic region accounting for measurement errors (in Russian). Proc. Arctic Ant. Res. Inst.,323, 40–74.

  • de la Mare, W. K., 1997: Abrupt mid-twentieth century decline in Antarctic sea ice extent from whaling records. Nature,389, 57–60.

  • Dickson, R., J. Lazier, J. Meincke, P. Rhines, and J. Swift, 1996: Long-term coordinated changes in the convective activity of the North Atlantic. Progress in Oceanography, Vol. 38, Pergamon Press, 241–295.

  • Greenspan, G., 1974: Discrete Numerical Methods in Physics and Engineering. Vol. 107, Mathematics in Science and Engineering, Academic Press, 312 pp.

  • IPCC, 1996: Climate Change 1995: The Science of Climate Change, J. T. Houghton et al., Eds., Intergovernmental Panel on Climate Change, Cambridge University Press, 572 pp.

  • Karl, T., 1998: Regional trends and variations of temperature and precipitation. The Regional Impacts of Climate Change: An Assessment of Vulnerability, R. T. Watson, M. C. Zinyowera, R. H. Moss, and D. J. Dokken, Eds., Intergovernmental Panel on Climate Change, Cambridge University Press, 412–425.

  • Kattsov, V. M., J. E. Walsh, A. Rinke, and K. Dethloff, 2000: Atmospheric climate models: Simulations of the Arctic Ocean fresh water budget components. The Freshwater Budget of the Arctic Ocean, E. L. Lewis, Ed., Kluwer Academic, in press.

  • Khrol, V. P., Ed., 1996: Atlas of Water Balance of the Northern Polar Area. Gidrometeoizdat, 81 pp.

  • Legates, D. R., and C. L. Willmott, 1990: Mean seasonal and spatial variability in gauge-corrected global precipitation. Int. J. Climatol.,10, 111–133.

  • Parker, D. E., and M. Jackson, 1995: The standard GISST data sets:Versions 1 and 2. Workshop on Simulations of the Climate of the Twentieth Century Using GISST, Bracknell, United Kingdom, Hadley Centre, 50–51.

  • Radionov, V. F., N. N. Bryazgin, and E. I. Alexandrov, 1997: The Snow Cover of the Arctic Basin. Applied Physics Laboratory Tech. Rep. APL-UW TR 9701, University of Washington, Seattle, WA, 95 pp. [Available from Applied Physics Laboratory, University of Washington, 1013 NE 40th Street, Seattle, WA 98105.].

  • Roeckner, E., and Coauthors, 1996: The atmospheric general circulation model ECHAM-4: Model description and simulation of present-day climate. Max Planck Institute for Meteorology Rep. 218, 90 pp. [Available from Max-Planck-Institut für Meteorologie, Bundesstr. 55, D-20146, Hamburg, Germany.].

  • Vinje, T., 1999: Barents Sea ice edge variations over the past 400 years. Proc. Workshop on Sea Ice Charts of the Arctic, Seattle, WA, World Climate Research Programme/Arctic Climate System Study, 4–6.

  • WCRP, 1996: Arctic Climate System Study (ACSYS): Initial implementation plan. World Climate Research Programme WCRP-85 WMO/TD-No. 627, 66 pp. [Available from International ACSYS Project Office, Polar Environmental Centre, N-9296 Tromso, Norway.].

  • Zakharov, V. F., 1997: Sea ice in the climate system. World Climate Research Programme/Arctic Climate System Study WMO/TD-No. 782, 80 pp.

Fig. 1.
Fig. 1.

Mean seasonal cycles and ranges of interannual variation during the period 1949–89 of area-averaged precipitation (mm day−1) for the ocean area poleward of 70°N simulated by ECHAM-4 (heavy solid line) relative to observational estimates of Bryazgin (1976) (B) and Legates and Willmott (1990) (L).

Citation: Journal of Climate 13, 8; 10.1175/1520-0442(2000)013<1362:TCTOAP>2.0.CO;2

Fig. 2.
Fig. 2.

Time series of annual precipitation for 55°–85°N: (a) ECHAM-4 simulated anomalies relative to 1961–90 model mean, and (b) observationally derived anomalies relative to 1961–90 observational means, as shown by IPCC (1996). Smooth curves are obtained from nine-point binomial filtering of annual values; thin curve in (a) is based on data for land only.

Citation: Journal of Climate 13, 8; 10.1175/1520-0442(2000)013<1362:TCTOAP>2.0.CO;2

Fig. 3.
Fig. 3.

As in Fig. 2 but for Karl’s (1998) “Arctic region”; observational data in (b) are from Karl (1998).

Citation: Journal of Climate 13, 8; 10.1175/1520-0442(2000)013<1362:TCTOAP>2.0.CO;2

Fig. 4.
Fig. 4.

As in Fig. 2b but based on a Laplacian interpolation in areas for which no observational data were available.

Citation: Journal of Climate 13, 8; 10.1175/1520-0442(2000)013<1362:TCTOAP>2.0.CO;2

Fig. 5.
Fig. 5.

Linear (least squares fit) trend of annual precipitation in ECHAM-4 model simulation. Trends are expressed as the percentage change (over 92 yr, 1903–94) relative to the 1961–90 mean at each grid point.

Citation: Journal of Climate 13, 8; 10.1175/1520-0442(2000)013<1362:TCTOAP>2.0.CO;2

Fig. 6.
Fig. 6.

Sea-ice extent anomalies, relative to 1903–94 mean, prescribed in the ECHAM-4 model simulation: (a) Northern Hemisphere and (b) Southern Hemisphere. Nine-point binomial filter has been applied to each time series.

Citation: Journal of Climate 13, 8; 10.1175/1520-0442(2000)013<1362:TCTOAP>2.0.CO;2

Fig. 7.
Fig. 7.

Sea surface temperature anomalies, relative to 1903–94 mean, in the ECHAM-4 model simulation: (a; top) Northern Hemisphere and (b; bottom) Southern Hemisphere. Nine-point binomial filter has been applied to each time series.

Citation: Journal of Climate 13, 8; 10.1175/1520-0442(2000)013<1362:TCTOAP>2.0.CO;2

Fig. 8.
Fig. 8.

Linear (least squares fit) trends of annual sea surface temperature in GISST dataset, 1903–94. Trends are expressed in °C per 92 yr.

Citation: Journal of Climate 13, 8; 10.1175/1520-0442(2000)013<1362:TCTOAP>2.0.CO;2

Save