Climate, Snow Cover, Microclimate, and Hydrology 55 



estimated average runoff (110 mm) from estimated average precipitation 

 (170 mm) gives a value of 60 mm yr"' for average annual evapotranspira- 

 tion. However, this estimate absorbs the inaccuracies of the other two 

 measurements and therefore has wide confidence intervals. In fact, other 

 data suggest that the actual value of evapotranspiration is substantially 

 higher. Annual Class-A pan evaporation in the Barrow area is about 160 

 mm (Brown et al. 1968); reducing this by a standard pan coefficient (0.6 

 to 0.7) gives a range of 96 to 112 mm yr"' for evaporation from a well- 

 watered surface. The energy balance data of Weller and Holmgren 

 (1974a) indicate evaporation rates of 4.8 mm day"' for the post-meh 

 period and 2.7 mm day"' during the summer season; if these rates are 

 considered average, an annual total of 210 mm is calculated. Stewart and 

 Rouse (1976) found that daily evapotranspiration from both wet and dry 

 tundra surfaces can be well estimated from net radiation and air temper- 

 ature. Application of their method using typical values for Barrow sug- 

 gests an annual total of about 140 mm. 



Interestingly, Stewart and Rouse (1976) found that evaporation 

 from a relatively dry tundra surface averaged 80% of that from the wet 

 surface (standing water) under the same temperature and radiation con- 

 ditions. This is apparently due to the fact that, as noted in Chapter 3, on- 

 ly 14 to 20% of the evapotranspiration from the land is due to transpira- 

 tion from vascular plants. The remainder is evapotranspiration from 

 mosses, which are often wetted by fog and dew and have low resistance 

 to water loss. 



These considerations therefore indicate that either the estimate of 

 regional average precipitation is too low, or the estimate of runoff is too 

 high, or both. It is likely that failure to account for occult precipitation 

 (fog, dew) is a significant source of error. In any case, it is important to 

 reahze that substantial uncertainties remain in our understanding of arc- 

 tic water balances, even in regions of relatively intensive study. 



Runoff is concentrated into a short period of time (Table 2-6). Al- 

 though the data are limited, there is a definite suggestion that runoff is 

 more time-concentrated in larger drainage basins. This is the opposite of 

 what would normally be expected, and may be due to the formation and 

 breakage of ice jams on the large streams. 



Actual data on infiltration are very limited, but it is possible to infer 

 the general nature of the intra-annual variation. In winter, an upward 

 moisture gradient is established, so that there is exfikration in the form 

 of vapor for much of the year. During and immediately after melt, water 

 infiltrates in liquid form, to the extent that soil moisture capacity is virtu- 

 ally reached. Through the summer, most of the water falling as precipita- 

 tion infiltrates, and most of this is subsequently evaporated and 

 transpired. 



Surface storage increases through the winter as snow accumulates 



