INTERNATIONAL JOURNAL OF CLIMATOLOGY Int. J. Climatol. 0: 000 – 000 (2016) Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/joc.4738 A study of the scaling properties of rainfall in spain and its appropriateness to generate intensity-duration-frequency curves from daily records Raúl Rodríguez-Solà , a M. Carmen Casas-Castillo, b * Xavier Navarro a and Ángel Redaño c AQ1 a Departament de Física, EPSEVG, Universitat Politècnica de Catalunya, BarcelonaTech (UPC), Vilanova i la Geltrú, Spain b Departament de Física, ESEIAAT, Universitat Politècnica de Catalunya, BarcelonaTech (UPC), Terrassa, Spain c Departament d’Astronomia i Meteorologia, Facultat de Física, Universitat de Barcelona, Barcelona, Spain ABSTRACT: A methodology based on the fractal properties of rainfall has been applied to obtain the intensity-duration-frequency, IDF, curves for 100 pluviometric Spanish stations over the Iberian Peninsula and the Balearic Islands from their daily precipitation series. The scaling behaviour of maximum rainfall intensities has been investigated and simple scaling has resulted suitable. This methodology has been veriied in three emblematic observatories with available sub-daily registers and current known generalized IDF relationships: the Fabra Observatory of Barcelona, the Ebre Obser- vatory near Tortosa (Tarragona) and the Retiro Observatory of Madrid. Despite some general concordance with the mean annual rainfall distribution over Spain, the spatial distribution of the scaling parameter found for the 100 stations shows some discrepancies in diverse areas probably due to the inluence of other features, as the inter-annual rainfall variability and the contribution of convective rainfall to total precipitation, on the characteristic rainfall pattern in these areas. KEY WORDS simple scaling; fractal analysis; rainfall intensity; intensity-duration-frequency curves Received 23 September 2015; Revised 8 March 2016; Accepted 13 March 2016 1. Introduction The calculation of the intensity-duration-frequency curves (IDF curves), which remains as an important tool for the risk analysis of natural hazards and hydrological design, usually requires a historical series of the maximum rainfall intensities at a sub-daily durations. Such in most locations rainfall data are usually available only from totaliser rain gauges registering 1-day precipitation, a method to infer intensity-frequency values for short durations from daily rainfall data can be very useful. There is a methodology based on the fractal properties of rainfall, or more specif- ically on the characteristic scale invariance of the fractal processes (Bendjoudi et al., 1997; De Michele et al., 2002; Yu et al., 2004), to obtain the disaggregation or down- scaling of low resolution precipitation data (daily) to high resolution (sub-daily) (Menabde et al., 1999; Desramaut, 2008; Bara et al., 2010). Many atmospheric processes pro- duced by complex dynamic mechanisms acting in a wide temporal range, as rainfall generation, give rise to phe- nomena that look the same regardless the scale where they are contemplated (self-similarity). These processes can be considered of fractal type and their properties exhibit power laws of the scale parameter , which is the ratio * Correspondence to: M. C. Casas-Castillo, Departament de Física, ESEIAAT, Universitat Politècnica de Catalunya, BarcelonaTech (UPC), Campus de Terrassa, Ediici TR2, C/Colom, 1, 08222 Terrassa, Spain. E-mail: m.carmen.casas@upc.edu t t 0 between any two durations t and t 0 within the scaling regime. The fractal self-similarity of natural processes, such as rainfall generation, has a statistical nature, as opposed to mathematical fractals in which the parties are an exact copy of the whole. Therefore the scaling proper- ties of phenomena like rain can be expressed by statistical relationships describing their fractal behaviour (Schertzer and Lovejoy, 1987; Schertzer and Lovejoy, 2011). Several studies analysed the rainfall process in Spain from a fractal point of view: Oñate Rubalcaba (1997) obtained a mean fractal dimension of 1.32 ± 0.01 for annual rainfall series recorded by 10 stations in the Iberian Peninsula, whereas Meseguer-Ruiz et al. (2014) compared the fractal dimen- sion of 10-min rainfall data from 20 Spanish observatories with a concentration index (CI) expressing the relative weight of the rainiest days of a series on the total accu- mulated rainfall of that series. Other studies investigated the multifractal character of rainfall records in Spanish stations (Valencia et al., 2010, 2015; García-Marín et al. 2008, 2013, 2015; Rodríguez et al., 2013). It has been widely observed that the probability distribution of the annual maximum precipitation inten- sities satisies scale relationships (Koutsoyiannis and Foufoula-Georgiu, 1993; Burlando and Rosso, 1996; Menabde et al. 1999), which means that the probability distribution of the annual maximum daily intensity I 24 (t 0 = 24 h) and the distribution at other scale I t can be related by a factor that is a power function of the scale parameter . In terms of the moments of order q, I q t , of the intensity 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 © 2016 Royal Meteorological Society