Author(s): Lucia Marotta; Vito Telesca
Linked Author(s):
Keywords: Weather Generator; Markov Chain; Climate Change; Wet and Dry Spells; Precipitation Amount
Abstract: The objective of this study is to analyse the capability of a Weather Generator based on a multivariate quasi-stationary and weakly depending stochastic process as a tool to take future decisions under the impact of a climate change. A Weather Generator, WG, is a statistical model to generate daily sequences of weather variables, such as precipitation, maximum and minimum temperatures and humidity. Among the different WGs available, there are those built in a two-step process: a first-order Markov chain to generate daily precipitation occurrence, and an exponential distribution to assign daily non-zero precipitation amounts up to a given threshold. ClimGen is widely used as a WG that belongs to this type. At this time, the purpose is to analyse if the most important hypothesis of ClimGen is reliable for the ⋅=) Mediterranean Region, or rather in the following expression the linear coefficient is constant and equal to. The parameter implemented in ClimGen is investigated for ALSIA and ECA stations close to the Mediterranean, particularly for daily precipitation series between 1959-2012. The results show the linear coefficient is not constant and it cannot be assumed as an average value for the analysed dataset because there is no correlation between the output data. The approaches implemented in ClimGen are rough (). The methodology has been tested at Policoro station (Basilicata Region, Southern Italy) for which a “new stochastic model”, that suits climate features including variability in frequency of wet days in a month, has been proposed to generate daily precipitation amounts. 75. 0=a 00. 1=a dwp (/ fa wet
Year: 2013