Experimental evaluation of the role of inorganic phosphorus for terrestrial carbon degradation in stream hyporheic zones

1. The hyporheic zone (HZ) is a hotspot of carbon processing in stream ecosystems as a consequence of the mixing of organic matter and nutrients from ground water and surface waters. However, major knowledge gaps exist regarding the drivers of microbial activity and carbon processing in the HZ among stream ecosystems with different carbon sources and sediment properties. We inves - tigated the impact of algal dissolved organic matter (DOM algal ) and inorganic phosphorus (P) on the degradation of soil DOM (DOM soil ) by hyporheic microor - ganisms in

tigated the impact of algal dissolved organic matter (DOM algal ) and inorganic phosphorus (P) on the degradation of soil DOM (DOM soil ) by hyporheic microorganisms in two laboratory experiments.
2. In our first experiment, we explored the influence of different ratios of DOM algal to DOM soil with and without P additions on microbial respiration and DOM composition in laboratory hyporheic microcosms under oxic conditions.Here, we used glass beads colonised by stream microorganisms resembling a pristine stream system.As the addition of DOM algal increased P concentrations, we added P to adjust the P concentrations to the same level of the pure DOM algal in a second batch.In our second experiment, we determined the aerobic microbial respiration of HZ-sediments from 20 streams along a land-use gradient in Austria incubated with DOM soil .Again, we performed the experiments with and without P additions to see whether effects on microbial respiration depended on the ambient P concentrations of the streams.
3. Aerobic microbial respiration in the hyporheic microcosms decreased with increasing DOM soil proportions.When P concentrations were adjusted to the P level of the DOM algal , aerobic microbial respiration rates were similar between the different DOM mixtures in the microcosms, mainly stimulating the degradation of humic-like DOM fractions.This highlights the stimulating effects of the P additions on hyporheic microbial respiration and humic-like DOM degradation in pristine streams.However, P additions caused a significant increase in microbial respiration in only one of 20 natural HZ-sediments, suggesting that

| INTRODUC TI ON
The mineralisation of dissolved organic matter (DOM) by microbes is an important biogeochemical process in running waters, which strongly influences nutrient and carbon cycling at the landscape scale (Battin et al., 2008;Marx et al., 2017).Microbial DOM processing in streams and rivers controls the rates of fluvial carbon retention and carbon dioxide (CO 2 ) outgassing, and alters the quantity and quality of the organic matter transported downstream (Boulton et al., 1998).Despite a growing number of studies on the microbial mineralisation and degradation of DOM in running waters (e.g., Fasching et al., 2014;Pucher et al., 2021), there are still large uncertainties regarding the mechanisms and factors determining the rates of fluvial DOM processing by microbes.
An important site for microbial DOM processing in streams and rivers is the hyporheic zone (HZ).The HZ is the sediment interface between surface water and ground water (Krause et al., 2017), and has been described as a hotspot of biogeochemical processing (Fasching et al., 2016;McClain et al., 2003).This property arises from the intense interactions between dissolved substances and reactive sites (i.e., particle surfaces and microbial biofilms) in the HZ and the mixing of qualitatively different surface water and ground water, promoting diverse microbial communities, metabolic pathways and chemical reactions (Boano et al., 2014;Findlay & Sobczak, 2000;Hedin et al., 1998;Nogaro et al., 2013).Steep redox gradients can form as a result of the high activity in this zone, further promoting the diversity of metabolic pathways and microbial communities (Krause et al., 2017).Studies have shown that the HZ can act as a sink for dissolved organic carbon (DOC) with ≤70% removal of groundwater DOC passing through this zone (Boodoo et al., 2019;Findlay et al., 1993).Rasilo et al. (2017) suggested that a major part of the terrestrially-derived organic carbon is mineralised in the HZ before entering surface waters of streams or rivers.Indeed, mesocosm and reach-scale studies have shown decreased DOC concentrations along flow paths through the HZ (Findlay et al., 1993(Findlay et al., , 2002;;Findlay & Sobczak, 1996;Schindler & Krabbenhoft, 1998).
The DOM occurring in the HZ consists of a broad spectrum of organic compounds of varying degradability and origin.In general, DOM refers to the total mass of the organic matter filtered through 0.45-0.7 μm filters, including carbon but also other elements present in the organic material, such as nitrogen, oxygen and hydrogen.
The main source of DOM in the HZ is soils (Billett et al., 2006;Caillon & Schelker, 2020;Marx et al., 2017), which is supplied mainly as ground water through the HZ via transient connection of hillslopes and riparian areas (McGlynn & McDonnell, 2003;Sawyer et al., 2014).The biodegradability of soil-derived DOM (DOM soil ) mostly has been described as low-to-moderate in various field and laboratory studies as a consequence of its aromatic molecular structures (Cincotta et al., 2019;Fellman et al., 2009;Hansen et al., 2016;Kalbitz et al., 2000), especially in comparison to DOM released from fresh leaf litter (Hongve et al., 2000).Furthermore, DOM soil generally is more oxidised, indicating that the relative energy per unit of carbon gained is lower (Del Giorgio & Cole, 1998) and the breakdown of this material requires more energy gained via respiration for the production of specific enzymes such as phenoloxidases (Berggren et al., 2012).Benthic algae also can leach copious amounts of DOM (Kaplan & Bott, 1989), which can be transported from the stream channel into the HZ (Wong & Williams, 2010).In contrast to DOM soil , algal DOM (DOM algal ) consists mainly of carbohydrates, amino acids and lipids with lower molecular weight compared to humic compounds (Wetzel, 2001).DOM algal usually is easily degradable and thus more susceptible to microbial processing than DOM soil (Hansen et al., 2016;Thorp & Delong, 2002).
In addition to DOM composition, enhanced inorganic nutrient concentrations in stream water, and here most notably the availability of inorganic phosphorus (P) (Elser et al., 2007), also may boost microbial activities and DOM degradation (Mutschlecner et al., 2018;Williams et al., 2010).The relevance of intrinsic properties of DOM defined by its molecular composition versus external factors (i.e., the physical, chemical and biological conditions, such as water temperature, redox state or externally supplied inorganic nutrients) for DOM degradation currently is widely debated (see, e.g., Catalán aerobic microbial respiration rates rarely were controlled by P availability in the investigated streams. 4. We conclude that nutrient pulses can, but do not necessarily, stimulate microbial activity and terrestrial carbon degradation in the HZ of streams.Nevertheless, at low ambient nutrient concentrations (i.e., in pristine streams) terrestrial carbon degradation in the HZ can be accelerated when nutrient pulses occur, which has consequences for CO 2 outgassing and the organic matter quality in the stream and its export to downstream sections.ecosystems.An experimental study with a standardised microbial community suggested that DOM composition could be more relevant for DOM degradation by planktonic microorganisms than the environmental conditions (Catalán et al., 2021).However, the role of P already present in lake and stream water has been shown to be especially relevant for the degradation of DOM sources with low degradability such as DOM soil (Fasching et al., 2014;Kragh et al., 2008;Williams et al., 2010).
Inorganic P can be supplied to an aquatic environment by anthropogenic sources, such as by fertiliser use from agriculture or by the release of treated or untreated wastewater.In nutrient-rich streams, the degradation of DOM soil in the HZ may be enhanced compared to pristine systems, leading to potentially increased hyporheic microbial respiration and also stream CO 2 outgassing.However, substantial knowledge gaps remain regarding the microbial respiration of different DOM sources (DOM soil and DOM algal = intrinsic control) in the HZ and the role of enhanced P concentrations (inorganic P additions = environmental control).
In this study, we aimed to experimentally analyse and quantify the aerobic respiration of DOM soil and DOM algal alone and in mixtures by hyporheic microbial communities, and to clarify the role of P in stimulating the respiration of DOM soil .For this purpose, closed model systems were used in laboratory experiments, which consisted of hyporheic sediment columns filled with colonised glass beads.In Experiment 1, we measured DOM degradation as microbial respiration rates under oxic conditions and the qualitative change of the DOM pool.We hypothesised that aerobic microbial respiration rates will be lower in DOM soil than in DOM algal and decrease with increasing fractions of DOM soil .This lower DOM soil availability will be compensated by P additions in our model systems.In Experiment 2, hyporheic sediments from streams along a P-pollution gradient were used to test whether any stimulating effects of P additions on the DOM soil degradation depended on the respective ambient P loads of the streams under oxic conditions.This second step also enabled us to compare our homogeneous model systems with the naturally occurring heterogeneous hyporheic sediments.Here, we hypothesised that aerobic microbial respiration of DOM soil will be higher at higher ambient P concentrations of the stream water.
By contrast, we expected the effects of P additions on microbial respiration to be more pronounced in streams with low ambient P concentrations.

| Preparation of soil and algal DOM extracts
We used forest soil sampled close to the Oberer Seebach (OSB), Lunz am See, Austria (47°51′08 N 15°03′59 E).The OSB is a pre-alpine second-order stream, draining a pristine, calcareous catchment of approximately 25 km 2 where vegetation is dominated by Fagus sylvatica and Picea abies.Average nutrient concentrations in the stream water are <4 μg L −1 soluble reactive phosphorus (SRP), 4.4 μg L −1 NH 4 -N, 570 μg L −1 NO 3 -N and 1.8 μg L −1 DOC (from Peter et al., 2014).The litter layer (O-horizon, ~1-2 cm depth) was removed and about 10 cm of the underlying topsoil (A-horizon, organic layer) was transported to the laboratory for the DOM soil extraction.The extraction of DOM soil was performed immediately after soil sampling.In the laboratory, the soil was sieved through a 4-mm sieve to remove coarse material, such as twigs or stones.Soil aliquots then were created by putting soil (500 g) into pre-combusted glass bottles (2 L; 450°C, 4 h).A soil slurry was created by adding ultrapure water (1.5 L; MilliQ, Millipore GmbH, Eschborn, Germany) and the slurry was shaken horizontally at room temperature in the dark for 48-72 hr.After the extraction phase, the slurry was centrifuged at 10,000 g (Avanti J-26 XP; rotor JA-14; 6 × 250 ml vessels; Beckman Coulter, Brea, CA, USA) and the supernatant was filtered through a pre-combusted GF/F filter (Whatman, nominal pore size 0.7 μm).
The DOM algal was extracted from Scenedesmus sp.cultures grown in WC medium (Guillard & Lorenzen, 1972) for 3-5 days.Scenedesmus spp.are common in any aquatic environment with high growing rates (Lürling, 2003) and, thus, are well-suited for this study.The algae were harvested by centrifugation, washed twice with ultrapure water to remove the medium and frozen to mechanically break the algal cells in a first step.After thawing, the algal suspension was ultrasonicated twice for 10 min with an ultrasonication tip (Sonifier W-250 D; Branson Ultrasonics, Danbury, CT, USA) at 50%.Similar to the soil, the algal suspension then was centrifuged and the supernatant filtered.Both sources were stored in the refrigerator and used within one to two days after extraction.

| Preparation and set-up of HZ microcosm incubations (Experiment 1)
We investigated the aerobic microbial respiration rates of DOM extracted from soils and algae at different mixing ratios, but similar carbon concentrations in vertical hyporheic sediment columns (Figure 1a,b).The columns consisted of acrylic glass cylinders (48 cm long, 4 cm diameter) filled with glass beads (soda-lime glass, diameter 3 mm; VWR International GmbH, Vienna, Austria), which were colonised with a natural microbial community from OSB.The glass bead surfaces were abraded in a shaker together with sand (AS200, Retsch GmbH, Haan, Germany; 5 min, amplitude 50) to facilitate the development of a biofilm.Thereafter, the glass beads were inoculated outside in a container continuously pumped through with water from OSB for 5 weeks.After the inoculation phase, the glass beads were transferred to the sediment columns and water from OSB was pumped through the cylinder in an up-flow circulating mode at a flow rate of 1 mL min −1 (Multichannel peristaltic pump 205S; Watson-Marlow Austria GmbH, Guntramsdorf, Austria) in the laboratory at 12°C in the dark for 10 days to acclimatise the For Experiment 1, DOM soil and DOM algal were added to the sediment columns alone (= pure end members; 100% DOM soil and 100% DOM algal ) and in mixtures with 10% increments of one end member and 10% decrements of the other end member (Figure 1b).This was done by first diluting the concentrated DOM algal and DOM soil extracts with ultrapure water to the same initial DOC concentration of approximately 10 mg L −1 .In a second step, the extracts were mixed in their respective percentages and equally diluted with OSB stream water (i.e., all received the same amount of OSB stream water), leading to 11 qualitatively different mixing regimes including the nine mixtures and two pure end members with comparable DOC levels (4.3 ± 0.6 mg L −1 ).Although DOC and dissolved inorganic nitrogen (DIN) concentrations were similar between mixtures, SRP concentrations varied greatly.Thus, in one set of microcosms, SRP concentrations were adjusted to the highest levels measured in the DOM algal source (~160 μg SRP L −1 ) by adding dipotassium hydrogen orthophosphate (K 2 HPO 4 ) and the other set of microcosms remained without SRP additions (Figure 1b).The two end members (100% DOM soil and 100% DOM algal ) were incubated in triplicate whereas the mixtures were not replicated.The microcosms were incubated in a climate chamber at 12°C in the dark for 10 days.

| Analyses of HZ microcosm incubations (Experiment 1)
At the start and end of Experiment 1, the following parameters were measured: concentrations of DOC, total dissolved nitrogen (TDN), NO 3 -N, NH 4 -N, NO 2 -N, total dissolved phosphorus (TDP) and SRP, as well as absorbance and fluorescence of DOM.For all analyses, about 100 mL water was needed.The water for the start measurements was taken directly from the bottles in which the mixtures were prepared before filling the sediment columns.The water for the end measurements, after 10 days, was sampled directly from the sediment columns.All samples for these chemical analyses were filtered over pre-combusted GF/F filter (Whatman; nominal pore size: 0.7 μm).Samples for TDN and TDP were digested before measuring by using persulfate in alkaline and acidic milieu, respectively (Eaton et al., 2005).Afterwards, samples for TDN, NO 3 -N, NH 4 -N, NO 2 -N, TDP and SRP were analysed on a continuous flow analyser (FlowSys 3rd generation; Alliance Instruments, Salzburg, Austria).NO 3 -N and NO 2 -N were analysed photometrically using the hydrazine reduction method to convert NO 3 -N to NO 2 -N (Kempers & Luft, 1988) and NH 4 -N using the indophenol blue method (Grasshoff & Johannsen, 1972).Furthermore, SRP was detected using the ascorbic acid colorimetric method (Hansen & Koroleff, 1999).The detection limit for NO 3 -N is 100 μg L −1 , for NO 2 -N 1 μg L −1 , for NH 4 -N 2 μg L −1 and for SRP 0.5 μg L −1 and uncertainty of the measurements for NO 3 -N is ±1.3%, for NO 2 -N ± 2.1%, for NH 4 -N ± 5% and for SRP ± 1.9%.The DOC concentrations were measured on a TOC Analyser (Sievers 900; GE Analytical Instruments, Boulder, CO, USA).The accuracy is given as <1% relative SD, and the precision is ±2% (GE Analytical Instruments).Reported C:N:P ratios for the two sources (DOM algal and DOM soil ) in this manuscript are based on GF/F filtered concentrations of DOC and TDN and TDP in molar units.
Concentrations of dissolved organic nitrogen (DON) are reported as the difference of TDN minus the sum of NO 3 -N, NH 4 -N and NO 2 -N.All glassware was acid-washed and combusted at 450°C for 4 hr before use.The quality of the DOM pool was assessed via absorbance on a Shimadzu UV-1700 spectrophotometer in 5-cm cuvettes with ultrapure water as blank and fluorescence via Excitation-Emission-Matrices with Hitachi Fluorescence Spectrophotometer F-7000 and 1-cm quartz cuvettes.Fluorescence intensities were measured at excitation (Ex) wavelengths ranging from 200 to 450 nm (5-nm increments) and emission (Em) wavelengths from 250 to 600 nm (2-nm increments).EEMs were corrected for blanks and the inner filter effect using corresponding absorbance measurements (Lakowicz, 2006).The area of the Raman peak of ultrapure water was used as a reference to express fluorescence intensities in Raman units (Lawaetz & Stedmon, 2009).
Respiration rates were measured in the microcosms via oxygen sensor spots (SP-PSt3-NAU; PreSens, Regensburg, Germany), allowing us to quantify changes in DOC inside the closed systems.
Oxygen was measured every ~24 hr during the incubations and oxygen concentrations were kept above 3.5 mg L −1 during all incubations to avoid changes in the redox state.Hence, the respiration rates that  we report here are aerobic respiration rates.Respiration was calculated from the decline (slope) of the oxygen concentrations over time during the first 6 days and presented as μgC L −1 day −1 , using 1 as a conversion factor for the respiratory quotient.However, we note that there can be large variations of respiratory quotients in fresh waters (see Berggren et al., 2012).As all microcosms had the same amount of glass beads and water, we calculated and here report the respiration rates per L pore water.Respiration measurements with no change of oxygen concentrations over the course of the measurements were regarded as zero (corresponds to an R 2 < 0.65; all three 100% soil and the 90% soil without P additions).

| Field sample incubations (Experiment 2)
In Experiment 2, we tested the effect of P additions on aerobic microbial respiration rates of DOM soil in hyporheic sediments of streams exposed to different ambient P concentrations.We selected 20 headwater streams from four regions in Austria to include a wide range of environmental conditions (for regions, ambient nutrient concentrations and days of sampling, see Table S3).
In each stream, the sediment was sampled with a spade from c. 10-20 cm depth after removing the upper layer and sieved through a 6.2-mm sieve onto a 2-mm sieve in the field retrieving the fraction between the two sizes (mean grain size 2-2.5 for all streams).
Stream water was collected for filling up the bottles for the incubations and concentrations of DOC, NO 3 -N, NH 4 -N and NO 2 -N, and SRP in the water was analysed as described before.Upon return to the laboratory, the HZ sediment was stored at 12°C in the dark until the start of the incubation.The incubations were set up within 1-6 days after sampling but all treatments for one stream always were incubated at the same time.For the incubations, subsamples from the HZ sediments (~30 g) were incubated in glass bottles (100 mL) with (1) DOM soil and (2) with DOM soil plus P mixed with the ambient stream water (Figure 1c).The DOM soil was freshly extracted for Experiment 2 after the procedure described above.We here used the same concentrations that we also used in Experiment 1 for the 100% soil incubations (~4-5 mg L −1 soil DOC and ~ 160 μg SRP L −1 ), leading to different starting DOC concentrations for each stream resulting from differences in the background ambient DOC.In addition, all HZ sediments were incubated in the respective pure stream water without any additions as control and to determine background respiratory activity (Figure 1c).Each treatment was incubated for c. 72 hr in four replicates.All incubations were done in the climate chamber at 12°C in the dark while mildly shaking the incubation bottles horizontally (~40 rpm, GFL 3019; Lauda-GFL, Burgwedel, Germany) for two days.The glass bottles had an oxygen sensor to measure oxygen decline over time using the same system as in the HZ sediment columns.Here, the oxygen concentrations were measured approximately every 12 hr and respiration was calculated as before from the slope of the oxygen concentrations over the whole incubation time, but given in μgC gDW −1 day −1 to account for slightly different weights of sediments in the bottles.After the incubations, the HZ sediment from each bottle was transferred to an aluminium bowl to determine the dry weight (DW) and organic carbon (OC) content for each bottle separately.
Additionally, subsamples for bacterial cell abundance, DW and OC content as ash-free dry mass (AFDM) as well as grain size analysis via sieving were taken from the original hyporheic sediment samples before the start of Experiment 2. The sediment DW and AFDM were determined after drying at 105°C and 450°C, respectively.Bacterial cell abundances were counted via flow cytometry after processing the sediment samples according to Duhamel and Jacquet (2006).Briefly, the fresh sediment samples were stored at −80°C until analysis.After thawing, the samples were diluted 200fold, stained with SYBR Green II stain (Invitrogen, Carlsbad, CA, USA), and analysed via flow cytometry (CytoFLEX; Beckman Coulter GmbH, Krefeld, Germany).

| Data analyses
Individual fluorescent components from the EEMs obtained from Experiment 1 were modelled using the drEEM MATLAB toolbox by Murphy et al. (2013).We set up a general model using all the EEMs from start and end that was validated by half-split validation.Next, several separate models (only start, only end without and with P additions, respectively) were set up to evaluate if some fluorescence components disappeared or appeared during the incubation (Table S1).The three models "START", "END+P" and "END" were compared using the "comPARAFAC" R-package by Parr et al. (2014).All four models showed five fluorescence components, of which three can be classified as humic-like and two as protein-like.Here, we report results from the general model including all data that were compared with the "openfluor" database (Murphy et al., 2014).In Table 1, the most closely associated references are presented.They were consistent with regards to origin: C1 (also often called peak T[ryptophan]) often is reported together with high aquatic primary production and algae leachate, whereas C2 (in literature often called peak M related to microbial or marine) is found in any environment, most probably because of its recalcitrant nature.Several studies on soil-DOM report a high occurrence of C3 and some the occurrence of C5 (also called peak A; see Coble, 1996).Tyrosine-or phenylalanine-like C4 has been found in dark environments and might be related to cell lysis.PARAFAC components are expressed as relative fluorescence intensities (ΣCi) using, % Ci = Ci � ΣCi × 100 %.Furthermore, the values for absorbance at 254 nm were extracted from the dataset, which has been used as a surrogate for terrestrial DOC concentrations (Brandstetter et al., 1996).
The following statistical analyses were conducted in R (v3.5.1) (R Core Team, 2018).For Experiment 1, simple linear regressions were used to assess the relationships between microbial respiration, changes in absorbance at 254 nm, sum of changes of the maximal fluorescence intensities of all five peaks (F max ), and the changes in relative intensities of all five identified peaks (C1-5) as dependent variables versus the fraction of DOM soil (in %) as an independent with DOM soil addition and mean respiration for the treatment (2) with DOM soil addition plus P additions (Spearman, 1910).These calculations were performed using the rcorr function of the Hmisc library in R on ln-transformed data (Harrell Jr & Dupont, 2006).

| DOM sources
The end members of Experiment 1, 100% DOM algal and DOM soil mixed with OSB stream water, were largely different in their nutrient and optical properties (Figure 2 and Figure S2; Table S2).The molar C:N:P ratio of the extracted DOM algal was 38:10:1, indicating that P most certainly was not limiting for C mineralisation according to the Redfield ratio (116:16:1; Redfield et al., 1963).The DOM soil had a molar C:N:P ratio of 226:68:1, indicating P limitation.The P addition changed the C:N:P ratio to 58:18:1 in DOM soil , thus reducing potential P limitations.Pure OSB water had a low absorbance and fluorescence compared to the extracted DOM sources, indicating that the optical signals in the pure incubations and mixtures were driven mostly by the two DOM extracts (Figure 2f-h).The relative fluorescence of tryptophane-like C1 was highest in 100% DOM algal (67%; Figure 2f; Table 1), whereas the other components were represented at much lower relative contributions with 7%, 6%, 10% and 10% for components C2 (humic-like and microbially degraded), C3 (terrestrial humic-like), C4 (tyrosine-like) and C5 (humic-like), respectively.

| Aerobic microbial respiration in HZ microcosm incubations (Experiment 1)
The aerobic microbial respiration rates in the HZ microcosms significantly decreased with increasing fractions of DOM soil at ambient P levels (Figure 3).The respiration rate in 100% DOM algal was 103 ± 10 μgC L −1 day −1 and decreased to zero with pure DOM soil as TA B L E 1 Fluorescence properties and general description of the five PARAFAC components identified by the global PARAFAC model for the HZ microcosm incubations from start and end of the incubations defined in the method section.The respiration rates in the P-adjusted incubations did not decrease with increasing DOM soil (Figure 3) and were slightly higher than the maximum respiration rate at ambient P levels (maximum measured for the mixture of 20% DOM soil with 80% DOM algal ).The aerobic microbial respiration for the P-adjusted incubations reached 145 ± 1 μgC L −1 day −1 when only DOM algal was added and 153 ± 16 μgC L −1 day −1 when only DOM soil was added.
An ANCOVA showed significant differences between the slopes of the regressions for incubations at ambient P versus adjusted P levels (p < 0.001).

| Changes in nutrients and optical properties in HZ microcosm incubations (Experiment 1)
DOC concentrations did not change substantially during Experiment 1 except in a few microcosms, where high DOC releases were observed (Figure 4a).Net changes in DOC concentrations were usually slightly negative in microcosms with >80% DOM algal , indicating net DOM uptake, and slightly positive with increased percentage of DOM soil , indicating net DOM release.Relative changes in coloured DOM (CDOM), expressed via the absorbance at 254 nm, were negative in all microcosms, with the largest decreases observed at 100% DOM algal and mixtures with a high content of DOM algal (Figure 4b).
In addition, DON (Figure 4c), TDN (Figure 4d) and TDP (Figure 4e) (Figure 4g).The humic-like component C5 showed a relative decrease across the whole gradient of DOM mixtures (Figure S2d) and tyrosine-like component C4, although only present in very low proportions at the start, increased relatively in all microcosms at both P levels (Figure S2c).

| Microbial respiration of field HZ sediments (Experiment 2)
The mean aerobic respiration of the field samples ranged from 0.24 μgC gDW −1 day −1 in Ebriachbach (Carinthia) to 4.72 μgC gDW −1 day −1 in Pfaffenbach (Styria) for the DOM soil incubations, and from 0.30 μgC gDW −1 day −1 in Großer Dürrenbach (Carinthia) to 4.18 μgC gDW −1 day −1 in Pfaffenbach (Styria) for the DOM soil incubations with P adjusted (Figure 5).The Wilcoxon rank sum test revealed significant effects of the P addition on the respiration rates only for the OSB stream (Table S3) (Wilcoxon W = 0; p = 0.029) (Figure 5a).Streams with lower ambient P levels from Lower Austria and Carinthia (0.4 to 2.6 μg L −1 SRP) showed a positive trend with a median increase in respiration of 0.12 μgC gDW −1 day −1 after P addition (Figure 5a).Streams with a higher nutrient level from Burgenland and Styria (8.3-22.0μg L −1 SRP) showed a negative trend with a decrease in respiration of 0.09 μgC gDW −1 day −1 after P addition (Figure 5b).Finally, we correlated respiration rates with and without P additions with the environmental variables DOC, NO 3 -N, NH 4 -N and SRP concentrations in the stream water, hyporheic bacterial cell abundances, and sediment OC content.The respiration rates were significantly positively correlated only with the sediment OC content, but not with DOC or inorganic nutrient concentrations in the water column (Figure 6).

| DISCUSS ION
In Experiment 1, we found negligible aerobic microbial respiration of pure DOM soil .However, higher microbial respiration rates were observed on DOM soil in the HZ microcosms after P was adjusted to P concentrations measured in the DOM algal , mainly stimulating the degradation of humic-like DOM.These results confirm that P limitation can be a key factor determining microbial activity and, hence, terrestrial DOM degradation under oxic conditions in pristine streams.However, in Experiment 2 across different streams with varying ambient nutrient concentrations, P addition did not significantly modulate aerobic microbial respiration rates.We detected a positive response of microbial respiration (mean increase) to P additions at low ambient nutrient levels and a negative response (mean decrease) at high ambient nutrient levels.Here, aerobic microbial respiration rates were only significantly positively correlated to sediment OC content.These findings highlight that increased supply of inorganic nutrients can boost HZ microbial communities, but this response is limited to pristine stream ecosystems where nutrients are low.At higher concentrations, factors other than inorganic nutrient supply drive microbial respiration such as sediment OC.Furthermore, interactions of different organic matter pools have been detected with important implications for carbon cycling in streams (Danger et al., 2013).More specifically, algal organic matter can fuel the mineralisation of more refractory organic matter  such as soil-derived DOM (Danger et al., 2013;Guenet et al., 2010Guenet et al., , 2014)).However, this priming effect (i.e., the stimulation of the mineralisation of refractory OM by the presence of labile OM; Guenet et al., 2010) is still debated in aquatic ecosystems (Bengtsson et al., 2018) and an experimental study in HZ microcosms did not detect a priming effect for leaf leachate (Bengtsson et al., 2014).We did not directly test if DOM algal fuelled the mineralisation of the refractory DOM soil because we did not explicitly distinguish the two end members and their carbon utilisation by the microorganisms, which is recommended when testing priming effects (Bengtsson et al., 2018;Guenet et al., 2010).Nevertheless, at ambient nutrient concentrations, the aerobic microbial respiration decreased significantly and linearly with increasing fractions of DOM soil .If DOM algal primes DOM soil mineralisation, we would expect a nonlinear response of microbial respiration with increasing DOM algal rather than the observed linear trend.This nonlinear response would indicate a possible non-additive effect by an additional mineralisation of the DOM soil with the introduction of the DOM algal .
In Experiment 1, we demonstrated that the addition of P increased aerobic microbial respiration in all sediment columns, independent of the ratio of DOM soil to DOM algal .This suggests that the microbial community in the HZ was limited by inorganic nutrient availability when fed with DOM soil under oxic conditions.Hence, the recalcitrance of DOM soil can depend on the stoichiometric balance of DOM soil as well as the recipient aquatic environment.The close link between DOC degradation and inorganic nutrient availability also has been shown by other studies focusing mainly on benthic biofilms (Wickland et al., 2012;Ziegler & Brisco, 2004).In a long-term nutrient addition study, Rosemond et al. (2015) found that inorganic nutrient enrichments reduced the residence times of stream water POC by about 50%.Particularly at downwelling sites, where oxic and nutrient-rich surface water infiltrates the HZ, hyporheic retention of terrestrial DOM can be high (Harjung, Perujo, et al., 2019;Krause et al., 2017).From an ecosystem perspective, elevated inorganic nutrient concentrations in streams may well enhance the processing of DOM soil in the HZ.The higher respiration then might contribute to CO 2 production and increased CO 2 concentrations in the stream that can be emitted (Boodoo et al., 2017;Peter et al., 2014).
We are aware that the molybden-blue method (Maruo et al., 2016) used here for SRP analysis may overestimate actual orthophosphate concentrations due to the hydrolysis of dissolved inorganic and organic P under the acidic environment of the analysis.
Furthermore, the bioavailability of different P pools, such as orthophosphate and dissolved organic P, may vary (Graeber et al., 2021;Soares et al., 2017;Stutter et al., 2018).These notions suggest that by adjusting the SRP concentrations of the DOM soil with additions of 100% bioavailable orthophosphate, we may have created gradients in the amounts of bioavailable P despite equal total P masses.
However, in Experiment 1, dissolved P concentrations decreased by 80-90% in all mixtures independent of whether the P originated from the DOM source or was added as orthophosphate to the treatments.This demonstrates that in our experiment both P sources were readily bioavailable to the microorganisms and we thus assume that there are no relevant implications of these observations on the experimental results.
We here focused on aerobic respiration and, hence, we cannot transfer our results to anoxic conditions commonly occurring in the HZ of fine-grained and/or nutrient-loaded streams.However, hyporheic redox conditions can be highly dynamic in space and time in all stream ecosystems (Kaufman et al., 2017) and the presence or absence of DOC and other electron acceptors may strongly influence hyporheic DOM soil degradation and microbial metabolism (Falkowski et al., 2008;Helton et al., 2015).Further research is needed to decipher the role of redox conditions on the microbial activities and degradation of terrestrial DOM in the HZ.

| No microbial stimulation by P addition in hyporheic stream sediments (Experiment 2)
In contrast to our expectations and also to the stimulating effects of P addition observed in the HZ microcosms experiment (Experiment 1), microbial respiration was not significantly increased by P additions in the sediments of the investigated streams (Experiment 2).
This lack of significant P stimulation was observed in all sediments except in OSB, independent of the respective ambient P concentrations in the stream water.However, we found a generally positive trend of increased respiration rates after P additions in streams with low P levels.This may indicate that the lack of significant P effects on microbial respiration may have been overlaid by other sediment characteristics such as organic matter or bacterial abundances.In our bottle incubations in Experiment 2, aerobic respiration rates significantly correlated with sediment OC content, pointing towards the importance of inherent controls of the HZ sediments on hyporheic DOM degradation.Additionally, sedimentbound but bioavailable P fractions may have masked the effect of the P provided in the water.Studies have shown that the activity and production of benthic epilithic biofilms is weakly linked to the external supply of DOC, N and P (Graeber et al., 2019;Kamjunke et al., 2015), probably as a result of increased internal organic matter and nutrient cycling and storage within the biofilms (Romaní et al., 2004).However, other studies clearly show a direct link between dissolved nutrients and the degradation of coarse particulate organic matter by the attached biofilm (Gulis et al., 2004;Stelzer et al., 2003).Interestingly, in streams with high water P levels, we observed a general (not significant) trend of decreased respiration rates after P additions.It is possible that the hyporheic microbial community in these nutrient polluted streams shifted its metabolism towards a higher assimilation of the provided DOM soil than respiration under P-enriched conditions.Similar effects have been found in Scandinavian lake waters where P additions increased bacterial growth and decreased cell-specific respiration at the same time (Allesson et al., 2020).However, more evidence is needed to confirm this mechanism here.Likewise, contrasting observations of tryptophan-like DOM lability can be found in the literature (Cory & Kaplan, 2012).It is possible that the tryptophan-like peak C1 represented two compositions with contrasting intrinsic bioavailability (e.g., containing DON or not) as protein peaks also have been associated with lignin phenols (Hernes et al., 2009).However, other external factors also could have controlled the C1-lability in our incubations (e.g., presence of high molecular weight carbon sources and inorganic nutrients).The latter is supported by the fact that DON did not decrease at adjusted P levels in soil incubations.
The proportion of humic-like C3 decreased more at higher DOM soil fractions and adjusted P levels as indicated by the significant linear relationship of ΔC3 with DOM soil at adjusted P levels.
We also observed a clear decrease in absorbance at 254 nm in all incubations, a region that usually is related to aromatic fractions of the DOM pool (Weishaar et al., 2003).Although the relative change of the absorbance at 254 nm (ΔA254) decreased with increasing DOM soil fractions, the incubations with high fractions of DOM soil had a higher A254 at the start of the experiment.This resulted in a higher absolute change at higher DOM soil fractions.Comparing ΔA254 of the 100% DOM soil , the mean relative decrease was higher at the adjusted (43%) than at the ambient (38%) P level.These results, in combination with the lower relative decrease in the tryptophanlike peak C1 and DON at high DOM soil fractions, suggest that the HZ microbes preferentially metabolise humic-like DOM soil when P is sufficiently available.
In all of our mixtures during the incubations, the relative proportion of the tyrosine-like component C4 increased or even appeared in the mixture dominated by DOM soil , suggesting that this component was released during microbial DOM degradation.These results provide evidence for the dual role of microbes as consumers and producers of fluorescent DOM (FDOM) in the HZ and thus support previous results that the HZ can be either a sink or a source of DOM (Battin et al., 2003).High FDOM production rates previously have been implicated in microbial degradation of natural DOM (Guillemette & del Giorgio, 2012;Lambert & Perga, 2019).
Interestingly, it mostly has been associated with microbes preferentially degrading algal-derived DOM in laboratory experiments (Rochelle-Newall & Fisher, 2002), and to natural systems where bacteria consume mostly algal carbon (Yamashita & Tanoue, 2004).
Although we found a high variability among the different HZ microcosms in Experiment 1, FDOM changes were not related to the algal to soil DOM gradient.This suggests that the FDOM production of tyrosine-like substances probably was not related to the respective DOM source, but may be explained by source-independent microbial cell lysis (Fox et al., 2017).Similar to the component C4, we found no clear patterns in the change of DOC concentrations (Figure 4a) or FDOM (Figure S2a) along the soil-algal DOM gradient.
Such non-conservative behaviour also was reported by Lambert and Perga (2019) who observed that DOM degradation of mixtures deviated considerably from predictions based on observations in the pure sources.
The EEM of the DOM soil was similar to the EEM of OSB, representing the natural fingerprint of the processed stream water.The humic-like and microbially degraded component C2 was highest in percentage in OSB and might be a photo-or biodegraded product of terrestrial humic-like C3 (Williams et al., 2010).Also, tyrosine-like C4 was contributing to the fluorescence mixture of OSB, suggesting that it represents mainly aquatic, degraded DOM.We therefore propose that the in-stream DOM represents a mixture of DOM soil and degraded DOM algal .

| CON CLUS ION
Microbial DOM removal in hyporheic sediments has been related to several abiotic factors such as sediment structure, hyporheic water flow and hydrological conditions, creating hot spots and hot moments of DOM degradation in the HZ during times and at locations of intense mixing, changing moisture content due to drying or gradients in oxygen availability (Fasching et al., 2016;Perujo et al., 2017;Reeder et al., 2018;Ruhala et al., 2018) biofilm, microbial respiration, nutrients, soil organic carbon 13652427, 2022, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/fwb.13980by CochraneAustria, Wiley Online Library on [09/05/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License et al., 2021; Kothawala et al., 2021).The fundamental question of what is of overall greater relevance, the DOM molecular composition or the surrounding aquatic environment with its dissolved nutrients, has not been comprehensively answered for oxygen-rich stream 13652427, 2022, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/fwb.13980by CochraneAustria, Wiley Online Library on [09/05/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License microbial community to laboratory conditions.The stream water was exchanged three times during that time.

F
Experimental setups for (a) laboratory flow-through microcosms, (b) the mixtures of algal and soil dissolved organic matter (DOM) for the gradient (both Experiment 1) and (c) field sample incubations (Experiment 2).
, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/fwb.13980by CochraneAustria, Wiley Online Library on [09/05/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Figure 1c) (Wilcoxon, 1945) because the data failed the Shapiro-Wilk test for normality.Additionally, pairwise Spearman's rank correlation coefficients (r s ) and their significance were calculated including nutrient concentrations of NO 3 − , NH 4 + , SRP and DOC directly measured in the ambient stream waters as well as hyporheic bacterial cell abundances, sediment OC content, mean respiration for the treatment (1) concentrations decreased in most HZ microcosms during the experiment.The incubations with adjusted P levels showed a decrease in TDN and TDP concentrations by c. 80-90% across all mixtures (Figure 4d,e).The relative DON changes were highest at 100% DOM algal and decreased with increasing soil fractions to almost zero at 100% DOM soil (Figure 4c).A different pattern was observed at ambient P concentrations.There, relative changes in DON, TDN and TDP closely followed the patterns observed in the incubations with adjusted P levels up to 70% DOM soil .Above this threshold, however, F I G U R E 2 Optical properties illustrated as excitation emission matrices (EEMs) of the components 1 to 5 (a-e) identified by the global PARAFAC model, the two end members (f) 100% algal dissolved organic matter (DOM) in Oberer Seebach (OSB) water and (g) 100% soil DOM in OSB water, and (h) pure OSB water at the start of the experiment, showing a typical EEM from OSB. Fluorescence intensities are in Raman units.In addition, (i) the relative contribution of the five peaks of each end member are shown.TDN and TDP showed much lower, whereas DON showed much higher decreases at ambient P concentrations than at adjusted P levels.The PARAFAC model revealed five fluorescence components that became either more or less pronounced in the microcosms during the incubations (Figure4f,g; FigureS2b-d).The fluorescence of DOM algal at the start was composed of almost 70% of tryptophane-like PARAFAC component C1.This high contribution to the fluorescing DOM portion substantially decreased during the incubations at both conditions, with and without adjusted P levels (Figure4f).The ANCOVA revealed significant differences between the two P levels regarding intercepts (p = 0.008), but not slopes (p = 0.912).This indicates a similar trend in the relative decrease in C1 for both conditions (R 2 = 0.95, p = 0.048 for ambient P levels and R 2 = 0.90, p = 0.02 for adjusted P levels).Conversely, DOM soil showed a fluorescence pattern where the typical algal component C1 accounted for only 24%, whereas the humic-like components C3 and C5 dominated (Figure2i).These two humic-like components including recalcitrant C2 behaved in a similar way across all mixtures during the incubations and neither slopes nor intercepts were significantly different (ANCOVA for slopes: p [C2] = 0.32; p [C3] = 0.37; p [C5] = 0.87 and for intercepts p [C2] = 0.54; p [C3] = 0.95; p [C5] = 0.46) (Figure 4g and Figure S2b,d).The terrestrial humic-like component C3 ranged from no changes in its relative abundance to small decreases with increasing DOM soil proportions

F
I G U R E 3 Microbial respiration in hyporheic zone (HZ) microcosms (in μgC L −1 day −1 ) at ambient (grey) and adjusted phosphorus concentrations (black).The dotted lines are the linear regressions of the microbial respiration with soil dissolved organic matter (DOM) fraction in % and the shaded areas denote the 95% confidence intervals.FI G U R E 4Relative changes in (a) dissolved organic carbon (DOC in mg L −1 ), (b) absorbance at 254 nm, and (c) dissolved organic nitrogen (DON), (d) total dissolved nitrogen (TDN) and (e) total dissolved phosphorus (TDP) (in %-change compared to the start) as well as changes in relative fluorescence intensities (% at end minus % at start) of (f) tryptophan-like C1 and (g) terrestrial humic-like C3.All data are displayed as difference from start to end of the HZ microcosm incubations at ambient (grey) and adjusted phosphorus concentrations (black).Positive values indicate an increase from start to end and negative values a decrease in intensities.Linear regression equations are only displayed when significant with p < 0.05 and the shaded areas denote the 95% confidence intervals of the linear regressions.No linear regression lines given for ΔDOC because of the outliers.4.1 | Drivers of DOM soil mineralisation in the HZ microcosms (Experiment 1)According to our expectations, aerobic microbial respiration was much lower in the HZ microcosms predominantly fed with DOM soil than in those with DOM algal without P additions.In fact, we did not measure a decrease in oxygen at 100% DOM soil , which suggests that the extracted DOM soil did not stimulate microbial activities.This is in line with studies from pelagic waters(Garcia et al., 2018;Hansen et al., 2016) and the oxic HZ(Wagner et al., 2014), which postulate that terrestrial DOM extracted from soils generally is less degraded in aquatic systems than DOM algal .By contrast,Fellman et al. (2009) found high proportions of bioavailable DOC in the pore water of different soils, often exceeding those of the stream water.Diverging results about the bioavailability of terrestrial DOM for aquatic microbial biofilms may be explained by differences in soil type, pore water residence time and soil bacterial demand for DOM, among others(Fellman et al., 2009;Tiefenbacher et al., 2020).Alternatively, the stability of soil DOC entering streams also may depend on the specific origin within the soil.For example,Cincotta et al. (2019) showed that DOC originating from upstream soil aggregates had significantly different properties to those from riparian aggregates.Moreover, the stability of soil aggregates, as well as the amount and molecular size of the DOC leached from aggregates depended on the ionic strength and Ca 2+ concentration of the water used to extract the DOC.Finally, during the 48-72 hr of soil leaching in the bottle before filtration and preparation for the incubations for our studies, a small part of the labile DOM soil already could have been degraded during the preparation (also reported inPucher et al., 2021).

F
I G U R E 6 Spearman's rank correlation (r s ) matrix on lntransformed data from the field sites and field sample incubations.Spearman's rank correlation coefficients (r p ) are given for each pairwise comparison but only significant correlations are displayed in colour (p < 0.05).R_soil and R_soilP = respiration rates of soil and soil plus phosphorus (P) incubations.F I G U R E 5Respiration rates of incubations with dissolved organic matter (DOM) soil (white boxplots) and DOM soil plus phosphorus (grey boxplots) from 20 Austrian streams divided into low ambient phosphorus level (a) with streams from Lower Austria (grey) and Carinthia (blue) and high ambient phosphorus level (b) with streams from Styria (brown) and Burgenland (green).The streams are sorted along a gradient of increasing phosphorus concentrations (on top of each stream in μg P L −1 ).At the right end, a boxplot (hatched) is shown with the data from differences between median respiration rates of DOM soil plus P treatments (data from grey boxplots) minus median respiration rates of DOM soil treatments (data from white boxplots) from each stream for low (a) and high (b) ambient P levels.The boxplots visualise the median of all stream sites (line), the first and third quartiles (hinges), and the 1.5 * inter-quartile ranges (whiskers).
13652427, 2022, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/fwb.13980by CochraneAustria, Wiley Online Library on [09/05/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 13652427, 2022, 11, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/fwb.13980by CochraneAustria, Wiley Online Library on [09/05/2023].See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions)on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 4.3 | Changes in DOM fluorescence patterns during incubations (Experiment 1) In Experiment 1, changes in microbial respiration rates as response to the two DOM sources were accompanied by alterations of the optical properties of the DOM.Mixtures dominated by DOM algal showed much lesser contribution of tryptophane-like component C1 at the end of the experiment, which was accompanied by high aerobic microbial respiration.This component C1 dominated the freshly leached DOM algal , but also accounted for a quarter of the soil fluorescence DOM pool at the start of the experiment.Hence, the freshly leached C1 from algae probably fuelled microbial activity, while C1 from DOM soil fractions decreased neither with nor without P addition compared to the other fluorescence components.

.
Our study contributes to unravelling the biogeochemical mechanisms of terrestrial DOM degradation in oxic hyporheic sediments.We provide experimental evidence that DOM soil degradation by hyporheic microbes can be stimulated by the addition of the limiting nutrient.Hence, external nutrient supply can become relevant in streams dominated by DOM soil of low intrinsic nutrient content.In these streams, nutrients are probably limiting the rates of microbial carbon processing and small changes in nutrient concentrations, caused for example by land-use change or wastewater inputs, can stimulate microbial respiration and terrestrial carbon degradation.However, our results from Experiment 2 with field samples suggest that the majority of the investigated HZ microbial communities most probably were not limited in P. To summarise, our results indicate that the strengths of the source or sink function of the HZ for DOM can vary depending on the origin of the DOM (terrestrial versus algal) and inorganic nutrient availability.More mechanistic studies are needed to look into the importance of organic matter and nutrient storage and cycling within biofilms compared to external supply to increase the understanding and prediction of DOM degradation in the HZ of different stream ecosystems.