Stream Dissolved Organic Matter in Permafrost Regions Shows Surprising Compositional Similarities but Negative Priming and Nutrient Effects

Abstract Permafrost degradation is delivering bioavailable dissolved organic matter (DOM) and inorganic nutrients to surface water networks. While these permafrost subsidies represent a small portion of total fluvial DOM and nutrient fluxes, they could influence food webs and net ecosystem carbon balance via priming or nutrient effects that destabilize background DOM. We investigated how addition of biolabile carbon (acetate) and inorganic nutrients (nitrogen and phosphorus) affected DOM decomposition with 28‐day incubations. We incubated late‐summer stream water from 23 locations nested in seven northern or high‐altitude regions in Asia, Europe, and North America. DOM loss ranged from 3% to 52%, showing a variety of longitudinal patterns within stream networks. DOM optical properties varied widely, but DOM showed compositional similarity based on Fourier transform ion cyclotron resonance mass spectrometry (FT‐ICR MS) analysis. Addition of acetate and nutrients decreased bulk DOM mineralization (i.e., negative priming), with more negative effects on biodegradable DOM but neutral or positive effects on stable DOM. Unexpectedly, acetate and nutrients triggered breakdown of colored DOM (CDOM), with median decreases of 1.6% in the control and 22% in the amended treatment. Additionally, the uptake of added acetate was strongly limited by nutrient availability across sites. These findings suggest that biolabile DOM and nutrients released from degrading permafrost may decrease background DOM mineralization but alter stoichiometry and light conditions in receiving waterbodies. We conclude that priming and nutrient effects are coupled in northern aquatic ecosystems and that quantifying two‐way interactions between DOM properties and environmental conditions could resolve conflicting observations about the drivers of DOM in permafrost zone waterways.


S1.1. Site Descriptions
The northern Alaska sites are longitudinally connected on the Sagavanirktok River, north of the Brooks Range in Alaska (NA). A shallow active layer limits the vegetation in this zone to plants with shallow rooting depths such as small flowering plants, grasses and low growing shrubs. This region is within the continuous permafrost zone and has high soil organic carbon (SOC), greater than 35 kg / m 2 (Grosse et al., 2011;Tarnocai et al., 2009).
The interior Alaska sites include two ice-rich silt-dominated sites including a low order tributary of the Yukon River (IA1) and a low order tributary of Hess Creek (IA2), and a mixed lithology headwaters catchment in the Beaver Creek drainage (IA3). The Yukon and Hess Creek sites are dominated by late Pleistocene to Holocene loess deposits (fine grained sediments, wind transported, and deposited and colluvially reworked in unglaciated areas) (Muhs & Budahn, 2006;Schirrmeister et al., 2013) with high syngenetic ice content, termed "yedoma" (French & Shur, 2010;Kanevskiy et al., 2011;Schirrmeister et al., 2013).
The northwestern Canada (NC) sites consist of three longitudinally connected sites along a tributary to and main stem of the Peel River in the Northwest Territories, Canada (NC1-3), with underlying substrate of fine glacial till hosting permafrost with high ice content. Thermokarst features in this region are characterized by large thaw-slumps with long debris tongues that deliver sediments and nutrients to downstream waterways (Kokelj et al., 2013;Littlefair et al., 2017).
The western Alaska sites (WA) are divided into two sub-regions (WA1-2 and WA3-4). WA 1 and WA2 are in an alpine setting within the Noatak National Preserve, an area managed by the U.S. National Park Service. These sites are underlain by coarse, glacially derived cobbles prone to draining freely. The textural nature of these sites does not promote permafrost growth or preservation nor high levels of organic matter storage and accumulation giving these waters a "pristine" character and low DOC concentrations. These two stream sites are located within the Agashashok River, which flows into the Noatak River. The other sub-region, WA3 and WA4 also lie within the Noatak National Preserve. WA3, the Imeleyak River flows downstream into the Cutler River (WA4) and ultimately into the Noatak. The WA3 and 4 sites are not comprised of the same coarse glacial alluvium as WA1 and 2. Rather these sites are underlain by permafrost and are more closely related to the northern Alaska sites (NA).
The northern Siberia (NS) sites are ice-rich and silt dominated (Dean et al., 2020;Weiss et al., 2016). Much of the syngenetic permafrost in Siberia was formed and deposited during the late Pleistocene in unglaciated regions of Siberia. However, these sites within the Indigirka River watershed drain younger Holocene materials and are dominated by thermokarst lakes and polygonal ice-wedge networks (Iwahana et al., 2014). These sites are in a low relief area and within proximity to one another (<1km). This small channel flows into the larger Indigirka River and eventually into the East Siberian Sea.
The Finland sites (FN) are located along the Simojoki River basin which traverses through a mosaic of peatlands (53%), conifer forests (40%), agricultural fields(3%), water (3%), and peat harvesting areas and urban developments (2%). These sampling locations represent a non-permafrost, peat environment and are much warmer receive more precipitation than the other sites. The uppermost sampling site is headwater system while the two other sites are located along the mainstem (24km from FN1 and 51km from FN2, respectively).

S1.2. Experimental Treatments and Additional Analyses
Stream water from each site was sampled in bulk (6 L), filtered (0.7 µm, Whatman GF/F precombusted 450 o C > 5hrs) into acid-washed 1 L amber bottles, and refrigerated until laboratory incubations were initiated. Experimental treatments (Table S.1) consisted of varying levels of labile carbon (acetate), which has been shown to be a substantial portion of DOC in permafrost ice cores (Stephanie A , and inorganic nutrients (NH4, NO3, PO4) known to be present in thaw waters S.A. Ewing et al., 2015;Wickland et al., 2018) and act as key electron donors and acceptors in soil-stream interfaces (Hedin et al., 1998). In addition to the solute subsamples, two 50 mL samples were collected in 60 mL Nalgene bottles at the initiation and termination of the incubation experiment (0 and 28 days) for Fourier Transform Ion Cyclotron Resonance Mass Spectrometry (FT-ICR MS) analysis of the character and composition of the organic matter in solution. FT-ICR-MS samples were frozen immediately and stored at -40˚C until analysis.

S1.3. Shimadzu V-TOC CSH
Analytical uncertainty for DOC and TN measurements are based on the reproducibility of known standards and working standards throughout the analysis procedure. Analyses were organized by site in order to achieve the best possible calibration range for increased accuracy, precision, and sensitivity.
Each study region was unique in the matrix effect (ambient solute concentrations) and could vary dramatically in terms of DOC and TN concentrations. Some regional groups that are not longitudinally connected expressed large differences in DOC concentrations between sites despite being regionally similar (IA1 DOC >> IA3 DOC), see Figure S.3.
Several standards were used to measure the uncertainty associated with each analysis. One consensus reference material (CRM) (Batch 11 Lot # 03-11 ) from the Hansell Laboratory at the University of Miami was used as reference for accuracy of our instruments relative to others who participated to the consensus. This CRM served as reference for both DOC and TN measurements, however, the concentrations present in this standard were much lower (0.5 mg C/L) than the DOC of several regions and often had an uncertainty up to 10% due to the high calibration range required for samples with acetate treatments. Standard_W, a well sample from Gallatin Valley, Montana was also used as a reference for both DOC and TN which is consistently 2.0 mg C/L and 3.0 mg N/L. A unique mixture of excess sample from the western Alaska sites was created to act as a representative standard for permafrost sites. This mixture was divided into two subsamples and one of which was spiked with acetate for a reference ion. Similar mixtures were prepared from excess northwestern Canada samples when the WA mixtures were depleted.
Experimental design and analytical methodology prioritized obtaining accurate acetate measurements over other solutes in solution. Acetate readily adheres to the column before most other anions in solution, however the acetate peak clusters together with other un-identified organics at the beginning of the chromatograph. To separate these peaks and precisely quantify acetate, the flow of carrier KOH (potassium hydroxide) is set at slow rate to separate the clustered peaks. KOH concentration was kept low (12mM) until the acetate had eluted and then the KOH concentration was ramped (to 39mM) to speed elution of other analytes.

Detailed protocol used by each regional team
Could priming and nutrient effects from degrading permafrost alter dissolved organic matter dynamics in permafrost-zone waterways?

Participants (region code)
Arctic Alaska ( It is unknown how hydrologic fluxes of organic matter will change in response to Arctic climate change, representing an important uncertainty in landscape carbon balance and habitat change in aquatic ecosystems (Laudon et al. 2012, Kicklighter et al. 2013, Abbott et al. 2016). Permafrost degradation is causing widespread release of biodegradable dissolved organic matter (BDOM) and inorganic nutrients (Vonk et al. 2013, Abbott et al. 2014, Treat et al. 2016. While permafrost BDOM is rapidly mineralized or diluted in headwater streams, , increased delivery of BDOM and nutrients to Arctic waterways could influence the turnover and mineralization of "modern" or background organic matter moving through freshwater and estuarine systems via priming and nutrient effects (Guenet et al. 2010, Rosemond et al. 2015. Because these lateral carbon fluxes are substantial (ca 120 tg yr -1 delivered to inland waters), priming of background DOM by permafrost BDOM or nutrients could influence landscape and continental-scale fluxes, but these interactions have not been quantified for Arctic or Boreal ecosystems (but see Dorado-García et al. 2015). This project will investigate the effects of BDOM and nutrients on DOM mineralization in headwater and mid-order streams. Sampling will occur during the growing season of 2016 in the following regions: Alaskan Tundra and Boreal Forest, Tibetan Plateau, Siberian Yedoma region, and Scandinavian boreal forest.
Priming and nutrient effects will be quantified by measuring DOC drawdown during incubations (Abbott et al. 2014 with and without biodegradable carbon (acetate) and nutrients. Sites will span a range of stream orders allowing us to quantify priming and nutrient effects in headwaters where permafrost effects will likely be greatest relative to background DOM fluxes, and in larger streams where currently the permafrost signal is washed out. Separation of added and ambient DOM losses will be determined by Bayesian inverse modelling with initial coefficients of decay constrained by incubations of the substrates (Hotchkiss et al. 2014), confirmed with direct quantification of acetate by ion chromatography. We will use optical properties (e.g. EEMs, CDOM, SUVA) and FTICR-MS sampling at the beginning and end of the incubation to investigate how DOM optical properties and composition determine initial biodegradability and what compounds were preferentially consumed. The study will be the first quantification of the priming effect in Arctic freshwater ecosystems and will address the conflicting evidence of the relationship between nutrient availability and DOM decomposition (Holmes et al. 2008, Wickland et al. 2012, Abbott et al. 2014.

Protocol
• Experimental design o We will quantify nutrient and priming effects moving from headwaters to larger rivers. For each region, samples should be collected from three nested catchments (including small headwater streams to larger rivers) to see how sensitive background or modern DOM is to priming and nutrient effects, and if this sensitivity changes during transport from land to sea.
• Sample collection o At each location, water should be filtered to 0.7 µm (GF/F) the same day as sampling and incubations should be set up within one to two days if possible. ▪ From each site collect: • 6 L of 0.7 µm-filtered water for setting up incubations.
• Three 30 mL samples of 0.2 µm-filtered water in amber vials for EEMs analysis in the Baker lab (refrigerated upon return to lab). • Three 60 mL of 0.2 µm-filtered water for background chemistry analyses (frozen upon return to lab). • Standard water chemistry parameters (estimate of discharge, conductivity, pH, O2 etc) and catchment characteristics (catchment size, permafrost extent, vegetation cover, wetland/peatland extent, basic climatological parameters, etc).
• Incubation setup o Incubations are run in 250 ml glass bottles at room temperature (20C). Remember that cleanliness is key throughout the experiment (ideally acid wash, ash, and rinse everything that can be prior to use) and avoid substances that can leach carbon (e.g. rubber caps). o At the beginning of the incubation, pour 200 mL of 0.7µm-filtered water into each glass bottle (filtering to 0.7µm removes particulates but lets enough microorganisms through to assure decomposition). o There are eight treatments run in triplicate for each site, two bottles for FTICR-MS from one site, and 5 blanks (bottles filled with DI) for each region (79 incubation bottles for each set of 3 sites). o Target acetate and nutrient concentrations are based on observed concentrations in thermokarst outflows (the likely upper limit of what soil water and headwater stream DOM would be exposed to). You will add one mL from the relevant stock solution (preparation described at end of protocol) for each treatment. ▪ Control (nothing added) ▪ 1, 5, and 10 mg/L of added acetate (CH3COONa) ▪ Low, medium, and high N/P treatments (NaNO3, NH4Cl, K2HPO4) ▪ 10 mg/L of Acetate + high N/P (1 mL from the high acetate and nutrient treatments) *Remember to pour off and freeze subsamples of each of your stock solutions and DI water into 15 mL Falcon tubes to be sent to Stephanie for analysis. This means 7 additional tubes.
o Because ~60 mL at the initial and final sampling are needed for FTICR-MS analysis (120 mL total), set up an additional "control" and "AN" (10 mg/L acetate and high N/P) bottle preferably for the headwater site. *Remember to pour off and freeze subsamples of each of your stock solutions and DI water into 15 mL Falcon tubes to be sent to Stephanie for analysis. This means 7 additional tubes.

Abbreviated protocol
1. Get water in field a. 6L for incubation b. Three 60 mL bottles filtered to 0.2 for "background" water chemistry (freeze) c. Three 40 mL amber vials filtered to 0.2 for EEMs analysis (refrigerate) 2. Set up incubation a. Filter the 6 L with 0.7 um glass fiber filter b. Pour 200 mL of water into 250 mL incubation bottles (for each site there should be 24 bottles: 8 treatments x 3 replicates). Additionally fill up 2 extra bottles for FTICR-MS (one to be run as a control and the other as a high acetate+nutrients treatment "AN") from your most interesting site (preferably the headwater site) and set up 5 blanks with DI water. c. Add 1 mL of treatment stock solution as appropriate (2 mL for acetate+nutrients treatment-one for each A and N ) 3. Samplings a. At 0, 7, and 28 days sample all the treatments (including blanks) i. Unfiltered water into 15 mL falcon tubes (3 tubes for acetate treatments, 2 tubes for all others