Effect of storm clustering on water balance estimates and its implications for climate impact assessment

Daily and monthly-based water balance computations are made for areas with climates ranging from humid (Coshocton, Ohio) through Mediterranean (Watsonville, California) and semi-arid (Dodge City, Kansas) to arid conditions (Tucson, Arizona). Monthly procedures lead to an underestimate of observed mean annual runoff by 14% in Coshocton, 59% in Tucson, and an overestimate by 9% in Watsonville. Daily balance calculations increase model accuracy. The improvement in runoff estimates by using the daily method is most significant for arid climates. Daily-monthly departures are greater in the semi-arid and arid areas than in the humid and Mediterranean areas. In terms of mean annual runoff, the difference between monthly estimates and daily estimates is 42.5% in arid Tucson, 58.2% in semi-arid Dodge City, but only 8.9% in humid Coshocton and 5.6% in Mediterranean Watsonville. The daily-monthly departures in soil moisture estimates are generally less than 10% in the humid and Mediterranean climates, but well above 50% in most months in the arid and semi-arid climates. Regression analysis indicates the daily-monthly difference in moisture surplus estimates correlates well with the amount of storm clustering within a month. Monthly computations depart increasingly from daily computations as storm clustering increases. The hydrological impacts of changes in storm clustering are studied by forcing the water balance model with daily precipitation sequences based on hypothetical storm scenarios. Total annual moisture surplus tends to increase with increased storm clustering. In the arid and semi-arid climates, the differences between the most and least clustering scenarios equal 35% up to 60% of surplus water generated by normal storms. They are about 20% in the cases of the humid and Mediterranean climates. These results suggest future potential changes in climatic variability such as storm delivery patterns can have significant impacts on water resource availability.