Nutrient loads exported from managed catchments reveal emergent biogeochemical stationarity

Abstract

Complexity of heterogeneous catchments poses challenges in predicting biogeochemical responses to human alterations and stochastic hydro-climatic drivers. Human interferences and climate change may have contributed to the demise of hydrologic stationarity, but our synthesis of a large body of observational data suggests that anthropogenic impacts have also resulted in the emergence of effective biogeochemical stationarity in managed catchments. Long-term monitoring data from the Mississippi-Atchafalaya River Basin (MARB) and the Baltic Sea Drainage Basin (BSDB) reveal that inter-annual variations in loads (L(T)) for total-N (TN) and total-P (TP), exported from a catchment are dominantly controlled by discharge (Q(T)) leading inevitably to temporal invariance of the annual, flow-weighted concentration, (C) over bar (f) = (L(T)/Q(T)). Emergence of this consistent pattern across diverse managed catchments is attributed to the anthropogenic legacy of accumulated nutrient sources generating memory, similar to ubiquitously present sources for geogenic constituents that also exhibit a linear L(T)-Q(T) relationship. These responses are characteristic of transport-limited systems. In contrast, in the absence of legacy sources in less-managed catchments, (C) over bar (f) values were highly variable and supply limited. We offer a theoretical explanation for the observed patterns at the event scale, and extend it to consider the stochastic nature of rainfall/flow patterns at annual scales. Our analysis suggests that: (1) expected inter-annual variations in L(T) can be robustly predicted given discharge variations arising from hydro-climatic or anthropogenic forcing, and (2) water-quality problems in receiving inland and coastal waters would persist until the accumulated storages of nutrients have been substantially depleted. The finding has notable implications on catchment management to mitigate adverse water-quality impacts, and on acceleration of global biogeochemical cycles. Citation: Basu, N. B., et al. (2010), Nutrient loads exported from managed catchments reveal emergent biogeochemical stationarity

Reconstructing historical changes in phosphorus inputs to rivers from point and nonpoint sources in a rapidly developing watershed in eastern China, 1980–2010

“…When significant changes in human activities occur over time, a long-term monitoring dataset should be divided into several multi-year time intervals for separately calibrating the LAM (Bowes et al, 2009a;Chen et al, 2013). Due to the difficulty in acquiring accurate long-term information on wastewater discharge for the large watershed in this study, the 31-yr time series was divided into several time intervals according to changes in the annual average flow-adjusted TP concentration (representing a net influence of human activities on PS and NPS pollution, Basu et al, 2010) and average flow-adjusted Cl − concentrations during the low flow regime (a proxy for the net influence of human activities on PS pollution, Jarvie et al, 2012;Ding et al, 2014). Due to the absence of a significant correlation between flow-adjusted concentrations and year number (p b 0.05) for a time interval, watershed human activities were assumed to have no significant influence on NPS and PS pollution dynamics within a given time period.…”

“…A linear regression analysis was adopted to determine the dependence of changes in annual flow-adjusted average TP and Cl − concentrations during the low flow regime versus year number (e.g., 1, 2, 3,…10) over selected annual time steps . Annual flow-adjusted average concentration (C FN , Basu et al, 2010) was estimated as C FN = ∑(C × Q) / Q a , where C is the measured concentration for a given monitoring date, Q is the corresponding river discharge for a given monitoring date, and Q a is average river discharge for all monitoring dates from 1980 to 2010. Regression and correlation analyses were performed using SPSS statistical software (version 16.0, SPSS Inc., Chicago, IL, USA).…”

“…Although P fertilizer application increased by 1.5-fold from 1980 to 1999 and then decreased by 37% from 2000 to 2010 (Supplementary material: Fig. S2b), NPS TP inputs continuously increased by~19-fold over the past 31 years, implying the increasing contribution of legacy P sources (Basu et al, 2010;Sharpley et al, 2013Sharpley et al, , 2015. This is supported by an average 2.0-fold increase in TP concentration (n = 209) and an average 1.4-fold increase in Olsen-P concentration (n = 406) measured in the upper 20-cm layer of agricultural soils between 1984 and 2009 in the Yongan River watershed (Chen et al, 2015).…”

“…The occurrence of legacy P sources implies that contemporary P source reductions would result in a limited decrease in NPS P input load to rivers (Meals et al, 2010;Sharpley et al, 2013). Therefore, adopting and improving interception measures (e.g., buffer strips, constructed wetlands) within the watershed would provide a more immediate reduction in NPS P loads compared to source reduction measures (Basu et al, 2010). Such interception measures are more pronounced for slowing water runoff and trapping sediment and P in the summer growing season.…”

“…Such temporal resolution is critical for application to the most sensitive times of the year when eutrophication is most likely to occur (Bowes et al, 2008(Bowes et al, , 2009a(Bowes et al, , 2009b. Although it cannot directly identify the individual contribution from specific nutrient sources (e.g., sewage treatment discharge, fertilizer application, animal and domestic wastes), an important advantage of the LAM is that it avoids the uncertainty derived from the lag effect of legacy phosphorus sources on NPS TP entering rivers, which has been widely observed in field-and watershed-scale studies (Basu et al, 2010;Sharpley et al, 2013Sharpley et al, , 2015Jarvie et al, 2013;Chen et al, 2015). The lag effect has not been strictly addressed in the majority of current watershed models, resulting in considerable uncertainty in estimating NPS pollution loads (Meals et al, 2010;Kleinman et al, 2011).…”