Persistence in daily and 5-day summer monsoon rainfall over India

Abstract

SummaryLinear regression equations are developed for predicting 10-day rainfall in each monsoon month (June through September) of 14 meteorological subdivisions lying mainly in Western, North-Western and Central India. The potential predictors considered are 5-day mean 700 mb level contour heights and their anomalies at 104 grid points covering an area between 10 ~ to 45~ and 20 ~ to 145~ and past 24 hour rainfall of the subdivision concerned. The 209 predictors are preliminarily screened by stagewise procedure and finally screened by forward selection procedure. The verification conducted on limited data of 12 cases revealed moderate success in June and September in predicting the 10-day rainfall of some subdivisions in 3 equally probable classes. The equations developed in the study are not suitable for operational use. Further work on this line is necessary before full potential of synoptic-statistical methods is realised. It is found that the contour heights over the Arabian sea and near Japan bear respectively significant negative and positive correlation coefficients with the 10-day rainfall of most of the subdivisions considered in all the months. Explanation of these relationships is attempted. Zusammenfassung Vorhersage zehntigiger Monmnniederschlfige fiber Indien

Markovian characteristics of dry spells over the Iberian Peninsula under present and future conditions using ESCENA ensemble of regional climate models

“…This means that the occurrence of a dry day depends on that the two preceding days had also been dry, implying at the same time persistence of the dryness. This first main result is in agreement with previous studies over different areas of the globe, such as India (Singh et al 1981) or a subregion of IB as Catalonia (Lana and Burgueño 1998), but it contrasts with those obtained by Martín-Vide and Gomez (1999) for the whole Spain. There, they conclude that the dry spells do not present a markovian behaviour in the South of Spain and the northern fringe needs a thirdorder to fit their empirical distribution to a Markov chain model.…”

“…General features of dry spells focused on their duration and frequency have been widely studied (Llano and Penalba 2011;Sarhadi and Heydarizadeh 2013;Serra et al 2013). However, some authors (Singh et al 1981;Anagnostopoulou et al 2003;Cindric ´ et al 2010) remark that a deeper analysis of the dry spells including the persistence, is needed to take into account also the temporal structure of the drought.…”

“…Different orders of model have been used to describe the phenomenon or tendency of dry (wet) periods to persist (Ochola and Kerkides 2003;Tolika and Maheras 2005;Schoof and Pryor 2008;Khoshghalb et al 2010;Mahmud and Deni 2011). However, different studies along twentieth century suggested a second order in the study of dry spells (Singh et al 1981;Lana and Burgueño 1998;Dastidar et al 2010). The translation of these studies into the future would allow to estimate how the dry spells characterization of a region would change as a response to climate change conditions.…”

Stochastic generation of annual, monthly and daily climate data: A review

“…Most Markov chain models referred in the literature are first order (lag one) (Gabriel and Newmann, 1962;Caskey, 1963;Weiss, 1964;Hopkins and Robillard, 1964;Bark, 1965, 1967;Lowry and Guthrie, 1968;Selvalingam and Miura, 1978;Stern, 1980a,b;Garbutt et al, 1981;Richardson, 1981;Stern and Coe, 1984). Models of second or higher orders have been studied by Chin (1977), Coe and Stern (1982), Gates and Tong (1976), Eidsvik (1980), Pegram (1980) and Singh et al (1981). The results varied with the climate characteristics of the rainfall stations investigated, with the statistical tests used and with the length of record.…”