干旱区生态环境知识资源中心

Seasonal distribution of rainfall is thought to be important in structuring arid and semi-arid plant communities. Controlled studies of rainfall distribution have proven difficult, especially at the scale necessary to evaluate plant community changes. Simply comparing years with different rainfall distributions is problematic because rainfall amounts are seldom constant, and it is impossible to factor out other climatic variables (such as temperature). The authors describe an approach for studying rainfall distribution using large (12 x 30 m) fixed-location rain shelters. Rainfall was excluded and water was applied to three zones within each of five individual rain shelters. The treatments applied to the watering zones were: (1) average precipitation distribution from long-term records (50% from November to March, 30% April to June, and 20% divided among July, August, and October); (2) spring distribution (80% from April to July); and (3) winter distribution (80% from November to March). This approach allows for a comparison of treatments where only rainfall distribution is altered. Factors to consider in designing a study of this nature include: cost of structures and plot size necessary, quality of water to be applied to plots; differential animal use of treatments; and management of watering treatments. (C) 1999 Academic Press.