Urbanization through the addition of impervious cover can alter catchment hydrology, often resulting in increased peak flows during floods. This phenomenon and the resulting impact on stream channel morphology is well documented in temperate climatic regions, but not well documented in the humid tropics where urbanization is rapidly occurring. This study investigates the long-term effects of urbanization on channel morphology in the humid sub-tropical region of Puerto Rico, an area characterized by frequent high-magnitude flows, and steep coarse-grained rivers. Grain size, low-flow channel roughness, and the hydraulic geometry of streams across a land-use gradient that ranges from pristine forest to high density urbanized catchments are compared. In areas that have been urbanized for several decades changes in channel features were measurable, but were smaller than those reported for comparable temperate streams. Decades of development has resulted in increased fine sediment and anthropogenic debris in urbanized catchments. Materials of anthropogenic origin comprise an average of 6% of the bed material in streams with catchments with 15% or greater impervious cover. At-a-station hydraulic geometry shows that velocity makes up a larger component of discharge for rural channels, while depth contributes a larger component of discharge in urban catchments. The average bank-full cross-sectional area of urbanized reaches was 1.5 times larger than comparable forested reaches, and less than the world average increase of 2.5. On average, stream width at bank-full height did not change with urbanization while the world average increase is 1.5 times. Overall, this study indicates that the morphologic changes that occur in response to urban runoff are less in channels that are already subject to frequent large magnitude storms. Furthermore, this study suggests that developing regions in the humid tropics shouldn’t rely on temperate analogues to determine the magnitude of impact of urbanization on stream morphology. Copyright © 2012 John Wiley & Sons, Ltd.

The hydrology and geochemistry of groundwater in tropical island aquifers, such as Barbados, Guam and Puerto Rico, are significantly influenced by tropical climatic conditions. Recharge to these aquifers is the product of regional and local climate patterns that control rainfall. Oxygen isotopes can be used to estimate the amount and timing of recharge on these islands because seasonal fluctuations of rainwater oxygen isotopic compositions are related to the amount of rainfall. The karst aquifers on Barbados, Guam and Puerto Rico have similar rainwater and groundwater oxygen isotopic compositions. Comparison of groundwater and rainwater oxygen isotopic compositions in the three aquifers indicates that: (1) recharge occurs by rapid infiltration with little evaporation prior to recharge; and (2) recharge is associated with similar monthly rainfall thresholds of 190–200 mm. These rainfall thresholds are remarkably similar for three aquifers in different geographic locations. Differences between the spatial variations of groundwater oxygen isotopic compositions on Barbados and Puerto Rico can be attributed to the more complex groundwater flow system on Puerto Rico. The surprising similarities of hydrologic conditions under which recharge will take place can be attributed to similarities in climate and geologic conditions, such as soils and limestone bedrock, that exist on the three islands. We therefore speculate that similar recharge-rainfall thresholds may be observed in other tropical karst aquifers.

From the hydrological point of view, mountains present somewhat of a paradox. Although they provide the bulk of the Earth’s freshwater resources, knowledge of the hydrological functioning of mountainous areas is generally much less extensive, reliable, and precise than that of other, often more easily accessible physiographic regions. Indeed, mountain regions have been referred to as ‘the blackest of black boxes in the hydrological cycle’ (Bandyopadhyay et al., 1997). Data collection networks are more difficult to set up and maintain in complex mountainous terrain, particularly in uninhabited forested headwater areas without road access, and minimum recommended instrumental densities are rarely met (Manley and Askew, 1993). Whilst the hydrological knowledge base on mountains in general has increased considerably in the last few decades, most montane research work has focused on determining catchment water and sediment outputs and their distribution in time and space; snow cover and glacier dynamics; or flood frequencies (Molnar, 1990; Lang and Musy, 1990; Bergmann et al., 1991; Young, 1992; Hofer, 1998), as opposed to the underlying hydrological processes (cf. Bonell, 1993). Until very recently (e.g. Motzer, 2003; Schellekens et al., 2004; Goller et al., 2005), the vast majority of this work dealt with mountains in the temperate zone, with very little pertaining to forested tropical mountains (see summaries of early research by Bruijnzeel and Proctor (1995) and Bruijnzeel (2001)). Knowledge of such processes would serve as a basis for increased understanding of how streamflows emanating from tropical mountains might change as a result of changes in climate, including the lifting condensation level, frequency and density of clouds and, by implication, water inputs and evaporative losses (Bruijnzeel, 2001). The average cloud condensation level on tropical islands can be as low as 600-800 m (Malkus, 1955), although on larger mountains situated further inland this may be closer to 2,000 m (Stadtmüller, 1987). Above this condensation level, the hydrology of the forest changes profoundly because of contributions of cloud water (i.e. fog) deposited to the forest canopy (Bruijnzeel, 2001). There is circumstantial evidence that complete conversion of these ‘tropical montane cloud forests’ (TMCF) to pasture or vegetable cropping may have an adverse effect on dry season flows or even on total water yield because of strongly diminished fog interception after clearing (Ingwersen, 1985; Brown et al., 1996). Similar effects may be expected when the average cloud condensation level is raised because of warming of the atmosphere due to global climate change (Still et al., 1999; Foster, 2001), or clearing of forest at lower elevations (Lawton et al., 2001; Van der Molen, 2002).