On the self-organization of hydrological fields: Fractal structure and long-term persistence

Land-atmosphere interactions play a fundamental role in the evolution of hydrological fields like rainfall, runoff and soil moisture. A water balance cellular automaton preserving the essence of such interactions and incorporating the basic hydrological equations inherent to the water balance was developed in this dissertation. According to our model, convective precipitation was a necessary condition to simulate scale-free behavior in total precipitation at a field level. However, convective rainfall should not be the dominant source of precipitation to avoid long-term antipersistence, a phenomenon not observed in nature. Synoptic processes incorporated in the model could explain fractal structure in soil moisture and evapotranspiration time series, but land-atmosphere interactions, such as convective rainfall or a feed back mechanism between soil moisture and precipitation, enhanced the scale-free behavior observed in these time series.; An uncorrelated geographical variation of the Hurst exponent was found in streamflow and precipitation records, using the Detrended Fluctuation Analysis. Long-term persistence was spatially distributed in different patterns for streamflow and precipitation time series, being more significant in the former than in the later. Feedback mechanisms and convective rainfall were suggested by our model as possible causes of the geographical variation of the Hurst exponent. Synoptic mechanisms associated with the soil storage process were also suggested as an additional explanation of the long-term persistence observed in streamflow records.; Finally, El Nino and La Nina could affect considerably the amount of precipitation at specific locations of the USA. However, the ENSO phenomenon did not disrupt significantly the structure observed in the rainfall. A methodology to determine the confidence intervals on the influence of El Nino and La Nina was developed. According to such a methodology, ENSO affected the amount of rainfall more significantly than its distribution.