Nutrient export and elemental stoichiometry in an urban tropical river

Abstract

Nutrient inputs to surface waters are particularly varied in urban areas, due to multiple nutrient sources and complex hydrologic pathways. Because of their close proximity to coastal waters, nutrient delivery from many urban areas can have profound impacts on coastal ecology. Relatively little is known about the temporal and spatial variability in stoichiometry of inorganic nutrients such as dissolved silica, nitrogen, and phosphorus (Si, N, and P) and dissolved organic matter in tropical urban environments. We examined nutrient stoichiometry of both inorganic nutrients and organic matter in an urban watershed in Puerto Rico served by municipal sanitary sewers and compared it to two nearby forested catchments using samples collected weekly from each river for 6 yr. Urbanization caused large increases in the concentration and flux of nitrogen and phosphorus (2‐ to 50‐fold), but surprisingly little change in N:P ratio. Concentrations of almost all major ions and dissolved silica were also significantly higher in the urban river than the wildland rivers. Yield of dissolved organic carbon (DOC) was not increased dramatically by urbanization, but the composition of dissolved organic matter shifted toward N‐rich material, with a larger increase in dissolved organic nitrogen (DON) than DOC. The molar ratio of DOC:DON was about 40 in rivers draining forested catchments but was only 10 in the urban river. Inclusion of Si in the assessment of urbanization's impacts reveals a large shift in the stoichiometry (Si:N and Si:P) of nutrient inputs. Because both Si concentrations and watershed exports are high in streams and rivers from many humid tropical catchments with siliceous bedrock, even the large increases in N and P exported from urban catchments result in delivery of Si, N, and P to coastal waters in stoichiometric ratios that are well in excess of the Si requirements of marine diatoms. Our data suggest that dissolved Si, often neglected in watershed biogeochemistry, should be included in studies of urban as well as less developed watersheds due to its potential significance for marine and lacustrine productivity.

Transport of N and P in U.S. streams and rivers differs with land use and between dissolved and particulate forms

“…Different land‐use practices and atmospheric inputs affect N:P ratio signatures in lakes (Arbuckle and Downing 2001, Elser et al 2009), but few studies have fully defined how pervasive nutrient loading from diverse watershed sources is manifested in terms of nutrient availability in streams, which may be at least as sensitive to human inputs (Maranger et al 2018). Contrary to the hypothesis that urbanization may increase P more than N (Duan et al 2012, McDowell et al 2019), we found both agriculture and urbanization increased dissolved and total N concentrations, suggesting that watersheds with intensive human activities are generally characterized by greater N loading relative to P. However, increases in N tended to be weaker in urban vs. agricultural watersheds. We also found that particulate P concentrations tended to be higher with greater urbanization, suggesting a stronger signal of particle‐derived P in these watersheds (e.g., possibly from diffuse sources such as increased soil or streambank erosion; [Withers and Jarvie 2008, Fox et al 2016] and/or particles influenced by treated or untreated sewage or septage).…”

“…In contrast, nutrient sources such as sewage effluent, runoff from urban watersheds, and amplified erosion of upland soils are likely to increase P relative to N (Downing and McCauley 1992, Withers and Jarvie 2008, Duan et al 2012). Although the predominant land use within a watershed is expected to affect stream N:P ratios, evidence for such shifts remains less explored in streams and rivers than in lake ecosystems (Downing and McCauley 1992, Vanni et al 2011, Ginger et al 2017, but see Maranger et al 2018, McDowell et al 2019). Distinct patterns of N and P loading likely prompt shifts in N vs. P limitation of critical ecosystem functions such as nutrient uptake, primary production and detrital breakdown (Rosemond et al, 2002, Tank and Dodds 2003, Dodds and Smith 2016), while also affecting the transport of N, P, and other elements downstream, eventually to coastal ecosystems.…”

“…Alternatively, we hypothesized that particulate nutrients would not track dissolved N:P ratios, reflecting watershed‐scale drivers that are independent of dissolved nutrient concentrations, and particulate N and P that are relatively inert and driven by abiotic factors. Overall, we expected that N and P concentrations and stoichiometry could inform patterns of nutrient sources related to land use, and so provide insight into processing mechanisms and transport as nutrients move downstream (McDowell et al 2019).…”

Arctic River Dissolved and Biogenic Silicon Exports—Current Conditions and Future Changes With Warming

“…Between 20% and 80% of DSi in soil solution has previously cycled through terrestrial vegetation (Alexandre et al, 1997;Clymans et al, 2016;Pokrovsky et al, 2013;Struyf & Conley, 2012;Zakharova et al, 2007), but DSi uptake by plants differs widely across species (Cornelis et al, 2010;Hodson et al, 2005). Thus, shifting vegetative cover alters rates of Si accumulation on terrestrial landscapes, which has been directly linked to altered riverine Si export from temperate (Carey & Fulweiler, 2012b;Conley et al, 2008;Struyf et al, 2010), subtropical (Chen et al, 2014), and tropical regions (McDowell et al, 2019). Standing water (e.g., lakes, ponds, wetlands) also retains Si on land, limiting export to downstream receiving waters (Humborg et al, 2000;Humborg et al, 2006).…”

Examining sources and pathways of phosphorus transfer in a ditch‐drained field

“…We used the loadflex package (version 1.1.11; Appling et al., 2015) in R (version 3.6.3) to estimate TDP loads in the drainage ditch at an hourly resolution as well as for the whole study period. Previously, researchers found the composite method (Aulenbach & Hooper, 2006) was most appropriate for load estimation in small watersheds (Coble et al., 2018; Kelly et al., 2019; McDowell et al., 2019) because it combines a regression‐based approach with interpolation. We also used the BaseflowSeparation() function in the EcoHydRology package (version 0.4.12.1; Fuka et al., 2018) in R (version 3.6.3) to separate hourly discharge into baseflow and stormflow components (see Supplemental Information).…”