EVALUATING AND EXTENDING CLIGEN PRECIPITATION GENERATION FOR THE LOESS PLATEAU OF CHINA

Climate generator (CLIGEN) is widely used in the United States to generate long-term climate scenarios for use with agricultural systems models. Its applicability needs to be evaluated for use in a new region or climate. The objectives were to: (1) evaluate the reproducibility of the latest version of CLIGEN v5.22564 in generating daily, monthly, and yearly precipitation depths at 12 stations, as well as storm patterns including storm duration (D), relative peak intensity (ip), and peak intensity (rp) at 10 stations dispersed across the Loess Plateau and (2) test whether an exponential distribution for generating D and a distribution-free approach for inducing desired rank correlation between precipitation depth and D can improve storm pattern generations. Mean absolute relative errors (MAREs) for simulating daily, monthly, annual, and annual maximum daily precipitation depth across all 12 stations were 3.5, 1.7, 1.7, and 5.0% for the mean and 5.0, 4.5, 13.0, and 13.6% for the standard deviations (SD), respectively. The model reproduced the distributions of monthly and annual precipitation depths well (p > 0.3), but the distribution of daily precipitation depth was less well produced. The first-order, two-state Markov chain algorithm was adequate for generating precipitation occurrence for the Loess Plateau of China; however, it underpredicted the longest dry periods. The CLIGEN-generated storm patterns poorly. It underpredicted mean and SD of D for storms >= 10 mm by -60.4 and -72.6%, respectively. Compared with D, ip, and rp were slightly better reproduced. The MAREs of mean and SD were 21.0 and 52.1% for ip, and 31.2 and 55.2% for rp, respectively. When an exponential distribution was used to generate D, MAREs were reduced to 2.6% for the mean and 7.8% for the SD. However, ip estimation became much worse with MAREs being 128.9% for the mean and 241.1% for the SD. Overall, storm pattern generation needs improvement. For better storm pattern generation for the region, precipitation depth, D, and rp may be generated correlatively using Copula methods.