Lacunarity, predictability and predictive instability of the daily pluviometric regime in the Iberian Peninsula

Abstract. The complexity of the daily pluviometric regime of the Iberian Peninsula is analysed from the point of view of its lacunarity, predictability and predictive instability. The database consists of daily pluviometric records obtained from 43 rain gauges in Spain and Portugal for the period 1950–1990. Five different series are generated for every rain gauge. The first series is constituted by the consecutive daily amounts. The other four consist of dry spell lengths with respect to daily amount thresholds of 0.1, 1.0, 5.0 and 10.0 mm/day. A dry spell length is defined as the number of consecutive days with rainfall amounts below one of these thresholds. The empirical lacunarity for every rain gauge is well reproduced by two power laws, the exponents varying notably from one gauge to another. The spatial distribution of the lacunarity is characterised by a north to south or southeast gradient, thus suggesting that this parameter can be a useful tool to distinguish between different pluviometric regimes. The predictability of the five series is quantified by means of the rescaled analysis and the interpretation of the Hurst exponent. Its patterns reveal that most part of the Iberian Peninsula shows signs of persistence for the daily rainfall and the dry spell series, although persistence is only clearly manifested in some small domains. The instability of possible predictive algorithms is analysed through the Lyapunov exponents. They are only computed for the series of daily amounts and for dry lengths respect to the threshold level of 0.1 mm/day due to the short number of dry spells for larger threshold levels. The series of daily amounts depict the highest instability along the Mediterranean coast. The series of dry spells show an increasing instability from NE to SW Spain, with a relevant nucleus of high Lyapunov values in the south-western Atlantic coast. As a summary, lacunarity and Hurst and Lyapunov exponents depict a relevant spatial variation, which is in agreement with well known patterns of the pluviometric regime, such as annual amount spatial distribution and return periods of dry spells.