The Metabolic Regimes at the Scale of an Entire Stream Network Unveiled Through Sensor Data and Machine Learning

Streams and rivers form dense networks that drain the terrestrial landscape and are relevant for biodiversity dynamics, ecosystem functioning, and transport and transformation of carbon. Yet, resolving in both space and time gross primary production (GPP), ecosystem respiration (ER) and net ecosystem production (NEP) at the scale of entire stream networks has been elusive so far. Here, combining Random Forest (RF) with time series of sensor data in 12 reach sites, we predicted annual regimes of GPP, ER, and NEP in 292 individual stream reaches and disclosed properties emerging from the network they form. We further predicted available light and thermal regimes for the entire network and expanded the library of stream metabolism predictors. We found that the annual network-scale metabolism was heterotrophic yet with a clear peak of autotrophy in spring. In agreement with the River Continuum Concept, small headwaters and larger downstream reaches contributed 16% and 60%, respectively, to the annual network-scale GPP. Our results suggest that ER rather than GPP drives the metabolic stability at the network scale, which is likely attributable to the buffering function of the streambed for ER, while GPP is more susceptible to flow-induced disturbance and fluctuations in light availability. Furthermore, we found large terrestrial subsidies fueling ER, pointing to an unexpectedly high network-scale level of heterotrophy, otherwise masked by simply considering reach-scale NEP estimations. Our machine learning approach sheds new light on the spatiotemporal dynamics of ecosystem metabolism at the network scale, which is a prerequisite to integrate aquatic and terrestrial carbon cycling at relevant scales.

Carbon and nitrogen dynamics in a soil profile: Model insights and application to a restored Swiss riparian area

David Andrew Barry, Jordi Batlle-Aguilar, Alessandro Brovelli, Pascal Niklaus

The key environmental importance of natural, healthy ecosystems has been progressively recognized and restoration of degraded lands towards their former natural state has become an area of active research worldwide. During restoration, environmental conditions (such as vegetation type and water availability) are manipulated to create ecological conditions suitable for the successful establishment of a target composition of species. Often, ecological restoration induces changes to adjacent ecosystems. This is the case of riparian ecosystems, and their restoration to their original undisturbed situation is likely to cause changes in nutrient cycles. For example, following the restoration of a riparian zone, microbial communities adapted to one set of environmental conditions have to acclimatize to another, and the subsequent changes in the composition of the biomass populations might induce changes in soil organic matter mineralization and soil respiration rates. Since the biogeochemical cycles are tightly interconnected, these changes can trigger nutrient storing or release, therefore inducing changes in nutrient cycles of adjacent ecosystems. Overall, the effects of the restoration activities on the hydrologic regime, soil properties and vegetation are still largely unknown and poorly understood. Within the RECORD project (http://www.cces.ethz.ch/projects/nature/Record), a large collaborative research effort undertaken to monitor and understand the changes in ecosystem functioning in riparian areas undergoing restoration, a numerical model has been developed to simulate the vertical transport of the mobile C and N components in a soil profile (model development discussed in the companion submitted abstract Batlle-Aguilar et al.). In the model, microbial decomposition of the soil organic matter drives biogeochemical transformations of C and N, while the activity of the soil biota is primarily controlled by the soil moisture content. The temporal evolution of the soil properties measured at one location of the RECORD experimental site, in a mixed forest dominated by ash and maple characteristic for the transition from riparian to upland forest, was used to validate the model and to gain insights into the key factors controlling the nutrient turnover. The site is located next to the Thur River where a revitalization project involving removal of levees has been implemented to create more natural conditions in the riparian zone. Soil water content and temperature at several depths were monitored continuously between October 2008 and October 2009. In October 2008, January 2009 and in biweekly frequency between April and October 2009, topsoil and soil solution at several depths were sampled. The soil solution samples were analysed for major carbon and nitrogen species, and the soil samples for denitrification enzyme activity, potential nitrification and related properties. In addition, soil respiration and N2O emissions were measured at each sampling event. Preliminary modelling results are shown, together with a discussion of the most influential parameters and processes controlling C and N turnover in riparian soils.

2010

Land use controls stream ecosystem metabolism by shifting dissolved organic matter and nutrient regimes

Amber Joy Ulseth

Stream ecosystem metabolism integrates production and respiration of organic matter and plays a fundamental role in the global carbon (C) cycle. Several studies have identified distal and proximal physical controls, for example, land use and transient storage, or the effects of water chemistry, that is, organic matter and nutrient availability, on stream metabolism. In parallel, research on organic matter quality has identified conspicuous gradients of chemical composition, yet mostly without demonstrating any functional implications. 2. We hypothesise that organic matter holds a key position in a more comprehensive causal framework of stream ecosystem metabolism, and that a concurrent study can improve mechanistic understanding. Specifically, we here postulate that dissolved organic matter (DOM) quality, that is, its chemical composition, acts as a control of ecosystem respiration (ER) as much as it is a result of gross primary production (GPP). As such, DOM quality likely forms a central link between land use and stream metabolism, besides known physical controls including transient storage and light availability. 3. To examine these hypotheses, we studied 33 streams in north-eastern Austria, a region with diverse land use ranging from semi-natural, forested areas to agricultural areas and settlements. We analysed DOM composition by absorbance and fluorescence spectroscopy, including modelling excitation–emission matrices with parallel factor analysis. We then opposed these data to GPP and ER estimated by fitting a metabolism model to single-station diurnal oxygen records. 4. Structural equation modelling revealed land use as a control on light conditions, DOM composition and concentration and nutrient concentrations, which together ultimately shaped GPP and ER. In particular, humified, coloured and aromatic DOM of predominantly terrestrial origin was prevalent in coniferous forest catchments and increased stream ER. Agricultural and urban areas enriched streams with phosphorous and nitrogen, which increased ER and GPP. Besides nutrients, GPP seemed to be weakly correlated with light availability and – in contrast to our hypothesis – left only a weak imprint on DOM composition. 5. Land-use change is rated as the most pervasive human influence on natural ecosystems and our results highlight its impact on aquatic GPP and ER in streams. To understand the role of inland waters in the global C cycle will require mechanistic understanding of ecosystem metabolism, which notably includes organic matter quality as a hitherto underappreciated key player.

Wiley-Blackwell2017

CLIMATE-INDUCED CHANGES IN SPRING SNOWMELT IMPACT ECOSYSTEM METABOLISM AND CARBON FLUXES IN AN ALPINE STREAM NETWORK

Tom Ian Battin, Enrico Bertuzzo, Amber Joy Ulseth

Although stream ecosystems are recognized as an important component of the global carbon cycle, the impact of climate-induced hydrological changes on carbon fluxes in stream networks remain unclear. We will discuss the effects of changes in snowmelt hydrology during the anomalously warm winter of 2013/2014 on gross primary production (GPP), ecosystem respiration (ER) and net ecosystem production (NEP) in an Alpine stream network. We estimated ecosystem metabolism across 12 study reaches of a 254 km2 subalpine catchment located in the Ybbs River Network (YRN), Austria, for 18 months. GPP peaked in 10 of our 12 study reaches during spring snowmelt, regardless of the amount of snow the previous winter. However, the shift from snow to rain precipitation following the low snow winter in 2013/2014 increased spring ER in upper elevation catchments, thereby shifting spring NEP from autotrophy to heterotrophy. Our findings suggest that the YRN evolved from transient sinks to sources of carbon dioxide (CO2) in spring as snowmelt hydrology differed following the high snow versus low snow winter. This shift towards increased heterotrophy during spring snowmelt following a warm winter has potential consequences towards annual ecosystem metabolism, as spring GPP contributed on average 33% to annual GPP fluxes compared to spring ER, which averaged 21% of annual ER fluxes. We propose that Alpine headwaters shunt less organic carbon to downstream ecosystems, but will emit more within-stream respiratory CO2 to the atmosphere annually, as the climate gets warmer.

2017