ABSTRACT
Karstic systems are highly heterogeneous geological formations characterized by a multiscale temporal and spatial hydrologic behavior with more or less localized temporal and spatial structures. Classical correlation and spectral analyses cannot take into account these properties. Therefore, it is proposed to introduce a new kind of transformation: the wavelet transform. Here we focus particularly on the use of wavelets to study temporal behavior of local precipitation and watershed runoffs from a part of the karstic system. In the first part of the paper, a brief mathematical overview of the continuous Morlet wavelet transform and of the multiresolution analysis is presented. An analogy with spectral analyses allows the introduction of concepts such as wavelet spectrum and cross-spectrum. In the second part, classical methods (spectral and correlation analyses) and wavelet transforms are applied and compared for daily rainfall rates and runoffs measured on a French karstic watershed (Pyrénées) over a period of 30 years. Different characteristic time scales of the rainfall and runoff processes are determined. These time scales are typically on the order of a few days for floods, but they also include significant half-year and one-year components and multi-annual components. The multiresolution cross-analysis also provides a new interpretation of the impulse response of the system. To conclude, wavelet transforms provide a valuable amount of information, which may be now taken into account in both temporal and spatially distributed karst modeling of precipitation and runoff.
Subject(s)
Disasters , Models, Theoretical , Soil , Water Movements , Forecasting , RainABSTRACT
Endemic goiter in Morocco is localized in the Rif and Atlas mountains. In this study, performed in the Skoura-Toundoute areas (high Atlas), we demonstrate that iodine deficiency is the main factor involved in the development of endemic goiter. However it may not constitute the only etiological factor. The median urinary iodine excretion, measured on untimed samples, was low both in goitrous patients, m = 18 micrograms/l (n = 109) and in non goitrous subjects, m = 24 micrograms/l (n = 47). The value found in a nonendemic region (Casablanca) was much higher m = 117 micrograms/l (n = 67). Another approach to evaluate iodine intake per day was the analysis of iodine content of some samples of cereals (barley, wheat and corn), drinking water, sea and rock salt, cow and breast milk. Iodine content is low, specially in the endemic area: barley = 90 micrograms/kg (n = 1); wheat = 227 micrograms/kg (n = 1); corn = 91 micrograms/kg (n = 1); water = 0.92 to 1.30 micrograms/l (n = 3); cow milk = 6 to 33 micrograms/l (n = 10) breast milk = 10 to 55 micrograms/l (n = 14); rock salt = 420 to 660 micrograms/kg (n = 8); sea salt = 90 to 160 micrograms/kg (n = 4). Sodium chloride intake per day, estimated by the measurement of sodium and chloride urinary excretion, is normal both in the Skoura-Toundoute areas (10.7 g and 9.8 g for goitrous and nongoitrous subjects, respectively) and in Casablanca (10.9 g for control persons). Altogether the results demonstrate the importance of introducing the practice of iodinated salt in the prophylaxis of endemic goiter in this area.
