An extreme decrease in dissolved oxygen concentrations resulting from litter decomposition in a mountain stream in the southern Kyushu Mountains [Japan]

The purpose of the present study was to clarify the influence of leaf litter decomposition on dissolved oxygen concentrations, one of the most important water quality indices, in mountain stream water in a mountain stream in the Southern Kyusyu Mountains. The following field measurements were conducted monthly from April 2005 to March 2006. Stream water samples were taken at each three site of springs, riffles, pools, and side-pools, and measured stream water temperature, electric conductivity and dissolved oxygen (DO) concentration (mg/l). The DO concentration was converted into DO saturation (%). The lengths, widths and depths of natural litter packs deposited in the pools and side-pools were measured and sampled. In the laboratory, each leaf pack was categorized into leaves (soft fraction), twig and bark (hard fraction) and organic fragments according to the extent of skeletonization and fragmentation and were dried at 80 deg C for 48 hours and weighed. The ratio of the weight of each category to the total weight of the sample was then determined and indices for estimating the degree of decomposition of each natural litter pack were developed. The results of multivariate analysis of variance with monthly change of mean DO saturation as an independent variable for each sampling sites and sampling season as the dependent indicated significant relationship without significant interactions evident between the dependent variables (P0.005) and Sheffe's multiple range test indicated a significant difference in the monthly change in mean DO saturation levels between riffles and side-pools (P0.05). Furthermore, in the pools and side-pools, the rapid decreases in DO saturation corresponded to the periods during which litter deposition was observed and a negative correlation was observed between the DO saturation and most of the litter decomposition indices. Correlations between the DO saturation and index-f (hard fraction) in pool 1, and between the DO saturation and index-a (soft fraction) in side-pools 2 and 3 were significant (P0.05). It was, therefore, inferred that not all of the litter deposited at a sample site was decomposed in the site, as most litter input from riparian forests into lotic systems is transported downstream while being decomposed and that the marked decrease in the DO saturations in relation to the litter decomposition should occur in a site satisfying the conditions for promoting litter retention such channel configuration, the area of the riparian zone, the hydrological and substrate characteristics along the stream margin, the discharge characteristics of the stream, biological conditions such as shredder density, and so on.