Spatial organization of stream water discharge and chemistry in forested headwaters

It has long been known that, in small catchments, stream water discharge and chemistry are highly variable but the variability decreases gradually with an increase in the catchment area. Wood et al. (1988) showed that model calculations of infiltration and the runoff rate became constant above a certain threshold area. They defined the threshold area as the representative elementary area (REA) and stated that above the REA only minimum knowledge of the underlying parameters is needed to explain the stream water discharge and chemistry. Subsequently, empirical studies were conducted in several catchments. These studies all verified the existence of an REA in real catchments and indicated that the REA values differed among catchments. The results also suggested that the confluence processes of stream water discharge and chemistry differed among catchments. However, it has not been clarified how the confluence processes behave and why processes differ among catchments. One of the unclear things to resolve is whether the variability of discharge and chemistry among small catchments can be regarded as randomness or if it is organized. Two previous studies examined it and reported the opposite results. Woods et al. (1995) reported that organization was apparent from their observations of specific discharge. However, Asano and Uchida (2010) stated that their results for SiO2 could be regarded as randomness. These studies targeted different observed items and different catchments. Therefore, general knowledge about organization of stream water discharge and chemistry has not been obtained. We observed spatial variability of stream water discharge and chemistry and examined the existence of spatial organization by using the statistical method. Our objective was to elucidate whether the spatial organization exists about stream water discharge and chemistry. Observations were conducted in three forested catchments in Japan. Snapshot samplings of stream water discharge and chemistry were conducted at low flow conditions. We compared observed data with the theoretical line suggested by Woods et al. (1995). Our results showed that the spatial variability in observed items did not completely decrease along the theoretical line in any catchment. Furthermore, the results for dissolved matter were different from the results for specific discharge. Specific discharge showed higher average values in small catchments than in the entire catchment, and convergence of specific discharge occurred more quickly than did convergence of dissolved components. These results meant that the variability and convergence of stream water discharge and chemistry couldn't be regarded as simple mixing of randomly distributed variability and verified the existence of organization. We discussed the factor of forming the spatial organization by using conceptual model considering micro topography.