Abstract. Riparian zones are habitats of critical conservation concern worldwide, as they are known to filter agricultural contaminants, buffer landscapes against erosion, and provide habitat for high numbers of species. Here we test the generality of the notion that riparian habitats harbor more species than adjacent upland habitats. Using previously published data collected from seven continents and including taxa ranging from Antarctic soil invertebrates to tropical rain forest lianas and primates, we show that riparian habitats do not harbor higher numbers of species, but rather support significantly different species pools altogether. In this way, riparian habitats increase regional (␥-) richness across the globe by Ͼ50%, on average. Thus conservation planners can easily increase the number of species protected in a regional portfolio by simply including a river within terrestrial biodiversity reserves. Our analysis also suggests numerous possible improvements for future studies of species richness gradients across riparian and upland habitats. First, Ͻ15% of the studies in our analysis included estimates of more than one taxonomic group of interest. Second, within a given taxonomic group, studies employed variable methodologies and sampling areas in pursuit of richness and turnover estimates. Future analyses of species richness patterns in watersheds should aim to include a more comprehensive suite of taxonomic groups and should measure richness at multiple spatial scales.

No abstract

1. Drylands worldwide are typified by extreme variability in hydrologic processes, which structures riparian communities at various temporal and spatial scales. One key question is how underlying differences in hydrology over the length of interrupted perennial rivers influence spatial and temporal patterns in species richness and species composition. 2. We examined effects of differences in dry season hydrology on species richness, composition and cover of herbaceous plant communities in the streamside zone (the zone influenced directly by low flows in the channel). Data were collected at ephemeral, intermittent and perennial flow reaches on three rivers of the desert Southwest (Arizona, U.S.A.): Lower Cienega Creek, Hassayampa River and Lower San Pedro River. 3. Patterns of species richness varied with temporal scale of analysis, that is between single-year and multi-year time frames. At the annual timescale, quadrat species richness (m )2 ) and herbaceous cover were higher at sites with perennial flow than at either intermittent or ephemeral sites. In contrast to this single-year pattern, the highest long-term richness occurred at intermittent sites. 4. Quadrat species richness, total species richness at a site (per 18 1-m 2 plots) and cover were more variable year to year at non-perennial sites than at perennial flow sites. On two of the three rivers, ephemeral sites had the highest inter-annual compositional variance, while the perennial sites had the lowest. 5. Compositional differences between the hydrologic site types were dominated by species turnover, not nestedness. The perennial sites had more wetland and perennial species than the other two site types. The intermittent sites had more annual species than did the other two types. 6. High long-term species richness and distinct species composition of intermittent sites are probably sustained by pronounced temporal variability in environmental conditions (i.e. frequent and persistent flow events, and dry periods). Plants at these sites take advantage of greater moisture than those at ephemeral sites and also experience less competition from resident species than those at perennial sites. 7. Conservation of desert riparian diversity depends upon the protection of consistently wet conditions at perennial flow sites, as well as the maintenance of the processes that cause fluctuations in environmental conditions at non-perennial sites.

The downstream effects of dams on riparian forests are strongly mediated by the character and magnitude of adjustment of the fluvial-geomorphic system. To examine the effects of flow regulation on sand-bed streams in eastern Colorado, we studied the riparian forest on three river segments, the dam-regulated South Fork Republican River downstream of Bonny Dam, the unregulated South Fork Republican River upstream of Bonny Dam, and the unregulated Arikaree River. Although Bonny Dam significantly reduced peak and mean discharge downstream since 1951, there was little difference in forest structure between the regulated and unregulated segments. On all river segments, the riparian forest was dominated by the native pioneer tree, Populus deltoides, which became established during a period of channel narrowing beginning after the 1935 flood of record and ending by 1965. The nonnative Elaeagnus angustifolia was present on all river segments, with recruitment ongoing. The lack of contrast in forest structure between regulated and unregulated reaches resulted primarily from the fact that no large floods occurred on any of the study segments since dam construction. Most of the riparian forest in the study area was located on the broad narrowing terrace, which was rarely inundated on the unregulated segments, resulting in little contrast with the regulated segment. A minor dam effect occurred on the small modern floodplain, which was actively disturbed on the unregulated segments, but not on the regulated segments. Although Bonny Dam had the potential to significantly influence downstream riparian ecosystems, this influence had not been expressed, and may never be if a large flood does not occur within the lifetime of the dam. Minor dam effects to riparian systems can be expected downstream of large dams in some settings, including the present example in which there was insufficient time for the dam effects to by fully expressed.