Page:Impact of Climate Change in 2030 Russia (2009).pdf/14

 latitude, annual precipitation has increased by about 4 percent over the past 50 years, especially over Russia’s permafrost-free zone and the entire Great Russian Plain. Over northern Russia, snow is providing a declining fraction of total annual precipitation.

The 20th century saw a trend of increased river output from the six largest Eurasian rivers flowing into the Arctic Ocean.xiii A similar trend is found in the climate simulation for the same period by the Hadley Centre's coupled climate model when the effects of manmade greenhouse gases are included. This finding is in line with predictions that global warming will cause changes in the water cycle.

Studies have shown that runoff in the Lena River increases in winter, spring, and (especially) the summer; and discharges decrease in autumn. These changes in seasonal streamflow characteristics indicate a hydrologic regime shift toward early snowmelt and higher summer streamflow, perhaps due to regional climate warming and permafrost degradation in the southern parts of Siberia. Winter snow accumulation is a major influence on summer and autumn discharge of the Ob and Yenisey Rivers and can affect winter and spring discharges of the Lena River, suggesting the importance of topography and permafrost conditions to river discharges in high-latitude regions.

Climate Predictions (Modeling)
Although Global Circulation (or Climate) Models (GCMs) can be used to infer climate changes in specific regions, developing models that have a high resolution sufficient to resolve local and regional scale changes is preferable. There are many challenges in reliably simulating and attributing observed temperature changes at regional and local scales. At these scales, natural climate variability can be relatively larger, making it harder to distinguish long-term changes expected due to external forcings.

The procedure of estimating the response at local scales based on results predicted at larger scales is known as “downscaling.” The two main methods for deriving information about the local climate are (1) dynamical downscaling (also referred to as “nested modeling” using “regional climate models” or “limited area models”), and (2) statistical downscaling (also referred to as “empirical” or “statistical-empirical” downscaling). Chemical composition models include the emission of gases and particles as inputs and simulate their chemical interactions; global transport by winds; and removal by rain, snow, and deposition to the earth’s surface.

Downscaled regional climate models rely on global models to provide boundary conditions and the radiative effect of well-mixed greenhouse gases for the region to be modeled. There are three primary approaches to numerical downscaling: (1) limited-area models, (2) stretched-grid models, and (3) uniformly high resolution atmospheric GCMs (AGCMs) or coupled atmosphere-ocean (-sea ice) GCMs (AOGCMs).

The magnitudes and patterns of the projected rainfall changes differ significantly among models, probably due to their coarse resolution. The Atlantic and Pacific Oceans are strongly influenced by natural variability occurring on decadal scales, but the Indian Ocean appears to be exhibiting a steady warming. Natural variability (from ENSO, for example) in ocean-atmosphere dynamics can lead to important differences in regional rates of surface-ocean warming that affect the atmospheric circulation and hence warming over land surfaces. Including sulfate aerosols in the models damps the regional climate sensitivity, but greenhouse warming still dominates the changes. Models that include emissions of short-lived radiatively active gases and particles suggest that future climate changes could significantly increase maximum ozone levels in already