Scaling and predicting solute transport processes in riverine ecosystems
Abstract
In the last three decades, research on solute transport and nutrient processing has revealed complex interactions between landscapes and stream ecosystems, and numerous attempts to scale and predict these processes have been primarily limited by the difficulty of measuring and extrapolating hydrodynamic and geomorphic characteristics. We hypothesize that there should be predictable patterns in the way that streams interact with their landscapes, because those interactions are in the form of energy, mass and momentum, which are conservative and interrelated properties. Therefore, despite local hydrogeomorphic characteristics define the actual extent of solute transport processes in a given riverine ecosystem, the physical imprints marked-up in breakthrough curves (BTCs) should have scaling properties. To evaluate our hypothesis we created an extensive database that includes 133 BTCs from conservative tracer experiments conducted under different hydrologic conditions (1 lt/s to 1197 m3/s), different experimental conditions (10s of meters to 10s of kilometers), different geographic positions (South and North America, Europe, Australia, Antarctica), and different types of lotic environments, i.e., urban manmade channels, forested headwater streams, desert-like streams, hyporheic wells, and major rivers. We investigated the existence of patterns in conservative solute transport using a model-independent approach, i.e., temporal moments of the histories of tracer experiments. Our results show that the normalized first absolute moment is correlated with the second and third moments with R2>0.99 for all riverine ecosystems. Most importantly, the first central temporal moment of the distributions (mean travel time) is correlated with the second (variance) with an R2>0.93, and the correlation between the second central moment and the third central moment (skewness) takes the form of the coefficient of skewness (CSK) with an R2>0.98, defining a statistically averaged CSK= 1.27 (1.07, 1.47), with 95% confidence bounds. These correlations were used to predict solute transport processes in four rivers in the UK (not included in the database used to develop the correlations) by parameterizing the Transient Storage model (R2>0.96 for all 4 rivers) and the Aggregated Dead Zone model (R2=0.87 for a channelized river, and R2>0.99 for the natural rivers) with the moment matching technique. Since our proposed technique to predict and scale solute transport processes has been developed from multi-ecosystem data around the world, and is only a function of discharge, length and travel times, i.e., quantities that are expected to appear in uncertainty analyzes, we propose its application as a fundamental routine to establish uncertainty bounds in decision-taking processes regarding solute transport in riverine ecosystems.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2012
- Bibcode:
- 2012AGUFM.H11E1239G
- Keywords:
-
- 0470 BIOGEOSCIENCES / Nutrients and nutrient cycling;
- 1813 HYDROLOGY / Eco-hydrology;
- 1830 HYDROLOGY / Groundwater/surface water interaction;
- 1872 HYDROLOGY / Time series analysis