The influence of native replanting on stream ecosystem metabolism in a degraded landscape: can a little vegetation go a long way?

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/263198253 The influence of native replanting on stream ecosystem metabolism in a degraded landscape: Can a little vegetation go a long... Article in Freshwater Biology · December 2013 DOI: 10.1111/fwb.12236 CITATIONS READS 9 57 4 authors: Darren Giling Mike Grace 15 PUBLICATIONS 71 CITATIONS 87 PUBLICATIONS 1,065 CITATIONS German Centre for Integrative Biodiversity Re… SEE PROFILE Monash University (Australia) SEE PROFILE Ralph Mac Nally Ross M Thompson 280 PUBLICATIONS 7,916 CITATIONS 139 PUBLICATIONS 2,441 CITATIONS University of Canberra SEE PROFILE University of Canberra SEE PROFILE Some of the authors of this publication are also working on these related projects: Lower Goulburn River Long-Term Intervention Monitoring Project View project Rede Amazônia Sustentável View project All content following this page was uploaded by Ross M Thompson on 01 August 2014. The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document and are linked to publications on ResearchGate, letting you access and read them immediately. Freshwater Biology (2013) doi:10.1111/fwb.12236 The influence of native replanting on stream ecosystem metabolism in a degraded landscape: can a little vegetation go a long way? DARREN P. GILING*, MICHAEL R. GRACE†, RALPH MAC NALLY* AND ROSS M. THOMPSON*,‡ *School of Biological Sciences, Monash University, Clayton, Vic., Australia † Water Studies Centre, School of Chemistry, Monash University, Clayton, Vic., Australia ‡ Institute for Applied Ecology, University of Canberra, Canberra, ACT, Australia SUMMARY 1. The effectiveness of revegetation is usually gauged by responses in biodiversity, which may differ between isolated replanted patches. The ecological value of revegetation may be detected more effectively by monitoring ecosystem processes. In-stream metabolism has been much modified by the degradation of riparian vegetation in agricultural landscapes around the world. We sought to determine whether the spatial scale typical of most riparian replanting projects (i.e. <1 km long) influences stream metabolism. 2. Metabolism is a functional measure that incorporates gross primary production (GPP), ecosystem respiration (ER) and the difference between them [net ecosystem productivity (NEP)]. We hypothesised that replanted reaches would have lower GPP (due to greater canopy shading) and greater ER (governed by larger terrestrial energy inputs) compared with pasture reaches. 3. We measured metabolism in paired reaches within four agricultural streams. Two streams had an untreated pasture reach contrasted with a reach replanted with native woody vegetation >17 years ago. The other two streams had similar riparian vegetation condition adjacent to both reaches, to act as reference sites. 4. Mean daily GPP (0.27–4.9 g O2 m 2 day 1) and ER (1.1–22 g O2 m 2 day 1) were within the range of values recorded previously in agricultural streams elsewhere. Replanted reaches had rates of NEP lower than upstream untreated reaches at treatment sites, but NEP was similar between reaches at reference sites. 5. The effects of replanting on stream ecosystem processes are detectable even when the spatial scale of restoration is relatively small in a whole-of-catchment context. Land managers can protect and restore vegetation at feasible spatial scales and benefit ecosystem processes. Ecosystem metabolism should be included in the range of responses that need to be monitored to provide a complete picture of the effectiveness of stream restoration. Keywords: agriculture, photosynthesis, respiration, restoration, riparian Introduction Human activity has pervasive effects on ecosystems, one of the most extensive being the conversion or removal of native vegetation for agriculture. Cropland and pasture now constitute the largest land-use category on the planet, and food production will have to increase to support a projected 50% increase in population by 2050 (Asner et al., 2004; U.S. Census Bureau, 2004). Landuse change has modified biogeochemical cycling and water availability and is a leading contributor to high extinction rates through habitat loss, fragmentation and degradation (Stoate et al., 2009). These land-use effects also affect adjacent stream ecosystems (Kominoski et al., in press); the loss of riparian vegetation reduces the supply of terrestrial organic matter and alters in-stream Correspondence: Darren P. Giling, School of Biological Sciences, Monash University, Building 18, Clayton, Vic., Australia. E-mail: darren.giling@monash.edu © 2013 John Wiley & Sons Ltd 1 2 D. P. Giling et al. biodiversity and food-web structure (Thompson & Townsend, 2004). Revegetation is intended to reverse effects of land-use change and to provide favourable ecological outcomes. However, many revegetation projects have ill-defined goals and are rarely monitored (Follstad Shah et al., 2007). When outcomes are assessed, success is often gauged by surveying biodiversity responses (Follstad Shah et al., 2007). However, responses in biodiversity may be highly variable or difficult to detect on short (years to decades) time scales (e.g. Parkyn et al., 2003; Munro, Lindenmayer & Fischer, 2007). An alternative approach to demonstrate ecological outcomes is to measure ecosystem processes. There has been a shift from activities aiming to restore biodiversity towards the restoration of entire ecosystems (Poiani et al., 2000). The latter includes ecosystem processes, their natural variability and the biodiversity that they support (Poiani et al., 2000). There have been several calls for a greater emphasis on the effects of restoration on ecosystem function (Bunn & Davies, 2000; Gessner & Chauvet, 2002), and even isolated patches of revegetated land in degraded landscapes may provide ‘ecosystem services’ (e.g. flood mitigation, nutrient cycling). For example, bioturbation by invertebrates increased soil water infiltration and restored local hydrological processes in a replanted open forest of Australia within 11–20 years (Colloff, Pullen & Cunningham, 2010). Monitoring the response of ecosystem processes to revegetation should be an important aspect of assessing replanting design and spatial arrangement. Riparian vegetation is important for stream ecosystems, providing terrestrial organic matter that subsidises stream metabolism (Roberts, Mulholland & Hill, 2007). Stream metabolism incorporates gross primary production (GPP; the production of organic carbon) and ecosystem respiration (ER; the consumption of organic carbon). The balance between these two processes can be expressed as a ratio GPP/ER (henceforth referred to as P/R) or as the difference between them, GPP – ER [net ecosystem production (NEP)]. This balance indicates whether a system is net heterotrophic (NEP < 0, more organic carbon is respired than is fixed) or net autotrophic (NEP > 0, surplus carbon is fixed and stored or exported) (Lovett, Cole & Pace, 2006). Natural forested stream ecosystems have NEP << 0 (e.g. Hagen et al., 2010), although secondary consumers may still obtain their energy needs from carbon produced in-stream (e.g. McCutchan & Lewis, 2002). The effect of agriculture on the rates of ER and GPP in streams has been assessed across forested agricultural land-use gradients (e.g. G€ ucker, Bo€echat & Giani, 2009; Young & Collier, 2009), but rarely used to assess the success of reach-scale restoration (but see Riley & Dodds, 2012). The primary determinants of ER and GPP are light, organic matter, nutrients and hydrology (Bernot et al., 2010). Of these, shading (reduced light supply) and the provision of terrestrial organic matter are expected to be modified by local replanting of riparian vegetation and changed towards values more characteristic of undisturbed systems; that is, lower NEP. Terrestrial organic matter is an important energy source and substratum supporting in-stream ER, while GPP can be light-limited (Young & Huryn, 1999; Mulholland et al., 2001). We aimed to determine whether the small scale (i.e. 100s of m long) typical of most riparian replanting projects is sufficient to affect stream ecosystem processes. Therefore, we measured whole-ecosystem metabolism in untreated and replanted reaches of four streams in degraded agricultural catchments of south-eastern Australia. We hypothesised that restored reaches would have a greater rate of ER, due to increased organic matter supply, and a decreased rate of GPP, due to greater shading. We assessed whether whole-stream metabolism could be a useful tool for monitoring restoration at reach scales over the mid-term (17–20 years), which is the age of many of the older replantings in this region and elsewhere. Methods Study design In-stream metabolism was measured in four lowgradient, second and third Strahler order (Table 1) agricultural streams in the Goulburn Broken catchment of south-eastern Australia. The dominant land use adjacent to the four study streams was dryland grazing by sheep and cattle. Stream substrata were dominated by gravel and sand. Comparisons were made between contiguous paired reaches on each stream (reach lengths, 210– 510 m). The riparian condition of each reach was classified as replanted (‘R’) or untreated and largely denuded of native vegetation (‘U’). Replanted reaches had been planted with Eucalyptus spp., Acacia spp. and Melaleuca spp. native to the area. The replantings were fenced, but livestock had occasional access. Untreated reaches were not fenced and had riparian vegetation typical of agricultural areas in the region, a ground cover of pasture grasses with occasional large remnant native trees (primarily Eucalyptus camaldulensis). The density of large © 2013 John Wiley & Sons Ltd, Freshwater Biology, doi: 10.1111/fwb.12236 © 2013 John Wiley & Sons Ltd, Freshwater Biology, doi: 10.1111/fwb.12236 Table 1 Physical, chemical and biotic site characteristics (mean  SD) for the eight study reaches measured during the 2012 data collection period Creightons Creek Honeysuckle Creek Warrenbayne Creek Moonee Creek Reach Up Down Up Down Up Down Up Down Treatment Untreated Untreated Replanted Replanted Untreated Replanted Untreated Replanted Parameter (unit) Travel time (min) K (day 1) Length (m) Width (n = 10) (m) Mean depth (n = 10) (m) Discharge (L s 1) Velocity (m s 1) Bed slope (n = 3) (cm m 1) Replanting age (years) Canopy cover (n = 5) (% closed) Surface PAR (mols m 2 day 1) Average temperature (°C) Turbidity (NTU) pH Electrical conductivity (lS cm 1) 1 NHþ 4 (n = 2) (mg L ) 1 NOx (n = 2) (mg L ) FRP (n = 2) (mg L 1) Total N (n = 2) (mg L 1) Total P (n = 2) (mg L 1) Dissolved organic C (n = 2) (mg L 1) Fluorescence index (n = 2) Total organic C (n = 2) (mg L 1) Fine particulate OM (n = 10) (g m 2) Coarse particulate OM (n = 15) (g m 2) Macrophyte cover (n = 10) (%) Water column Chl-a (n = 3) (ug L 1) Benthic Chl-a (n = 15) (mg m 2) 34 11 (3.0) 430 2.1 (2.1) 0.20 (0.09) 77 (2.7) 0.21 0.221 (0.001) – 21.1 (1.7) 34.6 (2.5) 13.6 (1.9) 5.6 7 0.144 0.022 0.305 0.017 0.71 0.065 5.7 1.46 6.6 160 (94) 11 (23) 13 (6.7) 2.3 (1.2) 5.6 (4.8) 31.7 7.1 (0.8) 380 2.4 (2.4) 0.19 (0.06) 110 10 (1.6) 510 5.8 (2.3) 0.37 (0.16) 59 (7.0) 0.08 0.468 (0.006) 12 69.1 (4.7) 12.8 (3.3) 15.9 (0.7) 19.9 7.36 0.164 0.011 0.077 0.024 0.77 0.095 9.4 1.44 9.8 670 (490) 250 (290) 14 (18.5) 1.2 (0.0) 0.6 (0.6) 0.2 0.215 (0.000) – 16.3 (0.9) 33.5 (2.7) 13.8 (1.9) 6.7 6.54 0.146 0.019 0.3 0.019 0.69 0.065 8.9 1.47 10 260 (230) 9.0 (20) 9.5 (5.1) 2.3 (0.0) 5.9 (9.2) FRP, filterable reactive phosphorus; PAR, photosynthetic active radiation. 129 5.2 (1.1) 540 5.6 (1.9) 0.37 (0.26) 0.07 0.298 (0.005) 12 62.2 (2.2) 21.4 (4.6) 16 (0.7) 31.5 7.28 0.164 0.014 0.091 0.024 0.78 0.095 9.7 1.44 9.8 740 (670) 300 (470) 19 (16) 1.2 (1.2) 24 (51) 23 52.2 16 (1.3) 6.8 (0.7) 210 250 4.8 (1.2) 5.3 (1.6) 0.41 (0.16) 0.39 (0.14) 145 (13) 0.15 0.08 0.110 (0.001) 0.032 (0.001) – 17 39.3 (4.2) 71.4 (3) 32.4 (11.5) 10.6 (3.7) 15.3 (2.7) 15.3 (2.7) 15.1 13.1 7.08 7.02 0.09 0.09 0.01 0.014 0.069 0.069 0.015 0.017 0.308 0.308 0.048 0.05 3.7 3.8 1.41 1.45 4 4.3 680 (340) 1100 (1000) 49 (86) 90 (120) 4.8 (4.3) 3.9 (3.9) 0.86 (1.0) 0.89 (1.0) 13 (12) 14 (20) 25.8 12 (3.4) 350 5.3 (0.8) 0.48 (0.08) 1058 (66) 0.23 0.358 (0.006) – 47.6 (1.6) 15.9 (0.4) 16.8 (0.5) 17.2 6.12 0.067 0.016 0.205 0.019 0.565 0.09 4.6 1.42 5 250 (160) 66 (150) 17 (13) 0.0 (0.0) 4.6 (5.2) 65.1 9.0 (4.4) 350 7.5 (1.3) 0.65 (0.14) 0.09 0.010 (0.000) 21 66.1 (3.8) 13.5 (1.1) 16.8 (0.5) 10.5 5.89 0.069 0.017 0.22 0.019 0.575 0.09 5 1.41 5.2 1800 (1300) 610 (1200) 2.4 (2.1) 0.38 (0.66) 2.5 (1.5) Isolated planting affects ecosystem function Site 3 4 D. P. Giling et al. remnant trees (in either U or R reaches) was similar between each pair of reaches (data not shown). Sites were not located close to any large areas of unmodified remnant forest. In two treatment streams (Warrenbayne Creek and Moonee Creek), the downstream reach was replanted 17 and 21 years ago, respectively, while the upstream reach was untreated (Fig. 1). These were termed the ‘untreated– replanted’ (UR) streams. The other two streams were selected to control for longitudinal metabolic variation, although there were no priori reasons to expect a difference in the rates of ER or GPP between reaches based on riparian condition. One stream (Honeysuckle Creek) had replanted trees (12 years old) in the riparian zone of both reaches and was termed the ‘replanted–replanted’ (RR) stream. The riparian zone of the fourth stream (Creightons Creek) had untreated riparian vegetation in both the upstream and downstream reaches, termed ‘UU’. There were no ‘reference’ lowland forested streams available in the study region for direct comparison. Our response variables were the mean difference in daily metabolic rate between reaches. The questions we posed in the Introduction can be stated formally as: ER and GPP UUdownstream UUupstream ¼ 0 ð1Þ ER and GPP RRdownstream RRupstream ¼ 0 ð2Þ ER URdownstream ð3Þ URupstream [ 0 GPP URdownstream ð4Þ URupstream \0 That is, metabolic rates between paired reaches at UU and RR will not differ, while replanted reaches at UR sites are hypothesised to have higher ER and lower GPP. UR Untreated/Replanted Upstream RR Replanted/Replanted U Downstream UU Untreated/Untreated U Creightons Creek R R Honeysuckle Creek U U R R Warrenbayne and Moonee Creeks Fig. 1 Design of metabolism experiment: grey boxes indicate replanted reaches (‘R’), and black circles show the location of dissolved oxygen loggers. Reach physicochemical characteristics Measurements of physical, chemical and biotic variables required to inform metabolic calculations or interpret results were made for each reach. Stream height was measured continuously using a TruTrack water level logger (Intech Instruments, Auckland, New Zealand). Stream width and mean water depth were measured at 10 haphazardly selected locations along each reach. Five evenly spaced hemispherical photographs were taken from the water surface mid-stream. Percentage canopy closure was estimated using Gap Light Analyzer software (version 2, Simon Fraser University, Burnaby, Canada). Spot measurements of pH, electrical conductivity (EC) and turbidity were taken using a U-50 Water Quality Meter (Horiba, Kyoto, Japan). Duplicate water samples were collected for measurements of total phosphorus (P) and total nitrogen (N) and analysed using the alkaline persulphate digestion method (APHA, 2005) using a Quick-Chem 8500 (Lachat Instruments, Loveland, CO, U.S.A.). Water was filtered onsite (0.45lm PES; Advantec, Dublin, CA, U.S.A.) for ammonium (NHþ 4 ), filterable reactive phosphorus (FRP) and nitrate plus nitrite (NOx). Concentrations of FRP, NHþ 4 and NOx were determined using flow injection analysis with the standard phosphomolybdenum blue, phenate and Griess methods, respectively (APHA, 2005). Water samples for total and dissolved organic carbon (DOC) were taken from mid-stream. Dissolved samples were filtered onsite through GF-75 glass fibre filters (Advantec) into pre-combusted, amber glass jars and acidified (to pH < 2) with concentrated (32%) HCl. These samples were refrigerated immediately for return to the laboratory, where they were split into two subsamples. For one subsample DOC concentration was analysed using a Shimadzu TOC-V CPH/CPN Total Organic Carbon Analyzer (Shimadzu, Tokyo, Japan) following APHA (2005) standard methods. The source (terrestrial or aquatic) of the dissolved organic matter (DOM) was determined by fluorescence spectrophotometry performed on the second DOC subsample. Fluorescence spectrophotometry characterises the complex mixture of DOM using the fluorescence index (FI; McKnight et al., 2001). FI was calculated as the ratio of emission intensity at 470 nm to emission intensity at 520 nm, at excitation wavelength of 370 nm on a Cary Eclipse fluorescence spectrophotometer (Varian, Melbourne, Australia) (McKnight et al., 2001). A FI of 1.3–1.5 indicates DOM originating from a terrestrial vegetation source, while FI of 1.7–1.9 indicates DOM of in-stream microbial origin (McKnight et al., 2001). © 2013 John Wiley & Sons Ltd, Freshwater Biology, doi: 10.1111/fwb.12236 Isolated planting affects ecosystem function Benthic coarse particulate organic matter (CPOM; >1 mm) was sampled by taking 15 cores from a range of water depths in each reach at haphazardly selected locations. A 25-cm-diameter core was inserted into the sediment, and all CPOM to a depth of 10 cm was removed and frozen. Samples were sieved, dried to a constant weight at 60 °C for 5–7 days and then ashed (4 h, 550 °C) to calculate ash-free dry mass (AFDM). Standing stock of benthic fine particulate organic matter (FPOM; 0.45 lm–1 mm) was estimated by taking 10 haphazardly selected cores in each reach by inserting a 7-cm-diameter core and removing all sediment and organic material. This material was sieved over nested 1-mm and 250-lm sieves. A well-mixed subsample of the material passing through the 250-lm sieve was filtered through a pre-ashed and weighed filter paper (Whatman GF/C). The sieved (>250 lm–1 mm) and residue (>0.45–250 lm) fractions were oven-dried and ashed, and the AFDM was summed to calculate total benthic FPOM AFDM. Water column and benthic chlorophyll concentrations were measured using the two-wavelength method (Nusch, 1980) with a UV-1700 UV-visible spectrophotometer (Shimadzu, Sydney, NSW, Australia). Triplicate water column chlorophyll samples were collected by filtering 500–1000 mL of stream water onto glass fibre filters (Whatman GF/C) and then freezing. Benthic chlorophyll samples were collected by taking 15 haphazardly selected 3-cm-diameter cores per reach. In soft sediments, a scraping of the top 3 mm of sediment was collected. On harder surfaces, a portion of the substratum was isolated and scrubbed before material was removed with a syringe and filtered. Chlorophyll was extracted from filter paper or sediment scraping with acetone (cold extraction at 4 °C for 12 h). Macrophyte (aquatic plant) areal cover was visually assessed at each width transect to the nearest 5%. Dominant macrophytes included Juncus spp., Persicaria spp. and Phragmites australis. Metabolism measurements Metabolism estimates were made over a single diel period at three streams in February 2011 (late summer) as a pilot study and subsequently over a longer deployment (6–16 days per site) in March/April 2012 (early autumn). We measured stream ecosystem metabolism using a whole-ecosystem, two-station approach (Odum, 1956), following a single-station analysis that enabled calculation of the reaeration coefficient (Atkinson et al., 2008). Dissolved oxygen (DO) and water temperature were 5 logged at 5-min intervals using D-Opto dissolved oxygen sensors (Zebra-Tech, Nelson, New Zealand). Probes were positioned mid-water column at three locations (up, mid and down; Fig. 1) on each stream to integrate DO change over the upstream (up-mid probe) and downstream (mid-down probe) reaches. Equipment limitations meant the reaches were contiguous (i.e. the middle oxygen probe was at the bottom of the upstream reach and at the top of the downstream reach). Measurements were made at each stream consecutively. Before and after placement, the loggers were put in an O2 saturated solution and then together in the stream for 1 hr to account for probe drift, and if required, linear corrections were applied prior to metabolism calculations. Photosynthetic active radiation (PAR) was measured at 5-min intervals using photosynthetic irradiance loggers (Odyssey, Christchurch, New Zealand). Light intensity was measured in an unshaded location and at the water surface in one location of the upstream and downstream reaches. Barometric pressure was logged with a Silva Atmospheric Data Centre Pro (Silva, Sollentuna, Sweden). Solute injection The average time taken for water to travel between the DO probes is used to estimate two-station areal rates of metabolism. Reach travel time was calculated using a solute slug of NaCl. Salt was dissolved and injected into a well-mixed area upstream of the probe location. EC was logged continuously at the reach boundaries using a HQd Portable Meter probe (Hach, Loveland, CO, U.S.A.) and a 90-FLT probe (TPS, Springwood, Qld, Australia) until EC returned to background values. The mean reach travel time was the time difference between half the salt passing the upstream and downstream probes. Calculation of metabolic rates Diel O2 and PAR data were used to estimate single-station rates of GPP, ER and the reaeration coefficient (K) for each DO probe on each day. We used Bayesian estimation to calculate the metabolic parameters in the daytime regression model (Kosinski, 1984):   D½O2 Ši p ¼ AIi R hðtempi tempave Þ Dt   þ K 1:0241ðtempi tempave Þ Di :; where A = photosynthetic constant, I = surface PAR (lEs m 2 s 1), p = photo-saturation exponent, R = respiration rate (mg L 1day 1), K = reaeration coefficient © 2013 John Wiley & Sons Ltd, Freshwater Biology, doi: 10.1111/fwb.12236 6 D. P. Giling et al. (day 1), D = O2 saturation deficit or surfeit (mg L 1) and h = temperature dependence factor. Data sets resulting in non-convergence of model parameters or poor model fits (R2 < 0.8) between measured and modelled DO for the single-station method were discarded from two-station calculations. Two-station metabolic estimates were made at each reach for days with successful single-station DO model fits for all three probes (between 5 and 8 days per stream). Two-station calculations were made using the diel oxygen mass balance approach (following Marzolf, Mulholland & Steinman, 1994) in an Excel spreadsheet (initially provided by C. Fellows, Griffith University, Australia). Upstream and downstream diel O2 data, the reaeration coefficient (determined by daytime regression model described above), barometric pressure, temperature, reach travel time and discharge were used to calculate the metabolic flux for each 5-min interval. Daily ER was calculated as the sum of metabolic flux for all night-time intervals, plus a temperature-corrected estimate for daytime (PAR > 0.5 lmol m 2 s 1) intervals. Daily GPP was calculated as the sum of daytime metabolic flux less the estimated daytime ER. These rates were divided by the stream bed area to convert to areal units. Statistical analysis We analysed the effect of replanting treatment (i.e. UR, RR or UU) using the difference between reaches (downstream – upstream) for each stream as the response variable to account for among-stream variation. A positive difference indicates that the variable was higher in the downstream reach. The effect of replanting treatment on predicted determinants of stream metabolism (i.e. benthic CPOM, benthic FPOM and canopy cover) between reaches was assessed with a two-object comparison (code available in Supporting Information). Multiple samples for each object provided means and variances for the comparison using WinBUGS (version 1.4; Lunn et al., 2000). Benthic CPOM mass and FPOM mass were log-transformed to improve distributional properties. Light at the stream surface was not included as a determinant because of low replication (one light intensity logger per reach). We analysed the effect of riparian treatment on metabolic variables for each stream and day (2012 data) using a linear mixed model (full code available in Supporting Information): responsei  Normalðli ; r2o Þ; ðlikelihoodÞ lij ¼ a þ b1 Ti þ b2 PARij þ b3 Tempij þ i ; ro  Uniformð0:001; 0:2Þ; a; bi  Normalð0; r2 ¼ 4Þ; ðpriorsÞ i  Normalð0; r2s Þ; rs  Uniformð0:001; 0:2Þ: Here, response is the daily difference in the metabolic variable between the downstream and upstream reaches, and treatment is the replanting category (UR, RR or UU). The model included covariance terms for total daily PAR and mean daily water temperature to account for differences due to non-concurrent measurements at different sites. A random effect for stream was included to account for the repeated daily sampling. We estimated treatment means and tested our hypotheses at the overall mean water temperature and mean total daily PAR. We excluded P/R from statistical analyses because ratios can be misleading and have poor distributional properties for statistical analysis (i.e. positively skewed and no upper boundary). We used the odds ratio to indicate an important effect of model parameters. The OR is the ratio of posterior odds to prior odds where 3 < OR < 10 indicates ‘substantial’ evidence and OR > 10 indicates ‘strong’ evidence of an effect (Jeffreys, 1961). An OR of infinity indicates that there is virtually no doubt that the parameter differs from zero given the uninformative prior (equally likely to be positive or negative). We used uninformative priors, so the prior odds were unity. Results Reach physicochemical characteristics Riparian vegetation affected the determinants of in-stream metabolism (i.e. light and organic matter availability). The difference in canopy closure between reaches was >0 (i.e. the canopy was more closed over the stream channel in replanted reaches) at UR sites (Fig. 2a; mean difference 25%  3.8 SD, OR = +infinity). The difference in canopy closure between up- and downstream reaches was not different from zero at the UU site (mean difference 4.8%  3.7 SD, OR = 9.4). The difference in canopy closure between reaches was less than zero (i.e. more open in downstream reach) at the RR site (mean difference 6.9%  5.0 SD, OR = 11), but was still much less than that between the reaches in the UR sites. The difference in benthic FPOM mass (AFDM g m 2) between reaches was >0 (i.e. greater in replanted reaches) at UR streams (Fig. 2b; mean difference © 2013 John Wiley & Sons Ltd, Freshwater Biology, doi: 10.1111/fwb.12236 Isolated planting affects ecosystem function 40 Δ % closed 30 (a) Canopy cover * 20 10 0 There was little variation in stream nutrient and organic carbon concentration between reaches in each stream (Table 1). Water temperature (measured by DO probes at 5 min intervals) also exhibited little variation among reaches (Table 1). FI results were also similar across all reaches and were consistent with DOM being from terrestrial sources (Table 1). * –20 Stream metabolic rates Δ AFDM g m–2 2500 2000 (b) FPOM (c) CPOM * 1500 1000 500 0 –500 1000 Δ AFDM g m–2 Mean (±SE) difference in downstream reach –10 7 800 Effect of riparian replanting on stream ecosystem metabolism 600 400 200 0 –200 UU There was large among-stream variation in mean daily two-station metabolism estimates (Table 2). Among all eight study reaches, daily GPP ranged from 0.06 to 5.7 g O2 m 2 day 1, and ER ranged from 0.38 to 27 g O2 m 2 day 1. Given that ER was generally greater than GPP, NEP was mostly negative and ranged from 25 to 0.21 g O2 m 2 day 1. Stream reaches were mostly heterotrophic, with only one reach on one day (from 65 reach-days) being autotrophic (P/R = 1.32; Warrenbayne Creek untreated reach). RR UR Treatment Fig. 2 Mean (SE) difference (downstream – upstream) in canopy closure (n = 5 hemispherical photographs per reach) and benthic organic matter standing stock between reaches at each of the four streams. Organic matter is split into the fine and coarse fractions (fine particulate organic matter and coarse particulate organic matter, n = 10 and 15 cores per reach, respectively). Positive differences indicate that the value was greater in the downstream reach. Asterisks indicate estimates different from zero at untreated-untreated (UU; white), replanted–replanted (RR; light grey) and untreated–replanted (UR; dark grey) from a linear mixed model at an odds ratio >10. 992  402 SD, OR = 19). The difference in benthic FPOM between reaches was not different from zero at the UU (mean difference 93.6  269 SD, OR = 2.0) or RR site (mean difference 74.3  515 SD, OR = 1.2). The difference in benthic CPOM mass (AFDM g m 2) between reaches was not greater than zero at UR streams (Fig. 2c; mean difference 292  88.2 SD, OR = 4.8). There was also no difference between reaches at the UU site (mean difference 1.50  29.7 SD, OR = 1.2) or RR site (mean difference 47.2  125 SD, OR = 1.4). The mean daily difference in metabolic rates (ER and GPP) between reaches was not different from zero at the UU site (Table 3; Fig. 3), in agreement with hypothesis 1. Similarly, there was no statistical difference in ER or GPP between reaches at the RR site (Table 3), in agreement with hypothesis 2. A mean 150% increase in ER rate was seen in replanted reaches compared with untreated reaches at UR sites (Table 2). Although not a strong effect in the overall model, the difference in mean daily ER between reaches was greater than zero at UR sites (Table 3; Fig. 3a), in agreement with hypothesis 3. The difference in GPP between reaches at UR sites was negative (i.e. smaller in replanted reaches), but the magnitude of this effect differed between the two sites. The mean daily difference in GPP between UR reaches was not different from zero (Table 3; Fig. 3b), refuting hypothesis 4. Riparian replanting affected NEP (Fig. 3c). There was no difference in mean daily NEP between reaches at UU or RR streams. In contrast, replanted reaches at UR streams had lower NEP than untreated reaches (Table 3; Fig. 3c). There were effects of the mean temperate covariate on the difference in ER and NEP between reaches (Table 3). Discussion Ecosystem respiration and GPP in our four agricultural streams in south-eastern Australia were within the range © 2013 John Wiley & Sons Ltd, Freshwater Biology, doi: 10.1111/fwb.12236 8 D. P. Giling et al. Table 2 Estimated metabolic rates (mean  SE g O2 m (R2 > 0.8) dissolved oxygen model fits 2 day 1) for diel periods during the 2011 and 2012 collection periods with successful Site Creightons Creek Honeysuckle Creek Warrenbayne Creek Moonee Creek Reach Up Down Up Down Up Down Up Down Treatment Untreated Untreated Replanted Replanted Untreated Replanted Untreated Replanted Summer 2011 Start date n days ER GPP NEP P/R 0 – – – – 0 – – – – 1 2.6 0.14 2.5 0.06 8 February 2011 1 4.5 0.49 4.0 0.11 1 6.8 2.6 4.2 0.38 24 February 2011 1 14.0 0.87 13.0 0.06 1* – – – – Autumn 2012 Start date n days ER GPP NEP P/R 7 1.9 0.55 1.4 0.38 5 April 2012 12 (0.42) 3.5 (0.07) 0.41 (0.37) 3.1 (0.09) 0.12 6 0.84 0.29 0.54 0.37 22 March 2012 6 (0.10) 1.3 (0.03) 0.25 (0.11) 1.0 (0.06) 0.20 9 1.7 1.2 0.47 0.79 29 March 2012 9 (0.24) 5.7 (0.11) 0.29 (0.19) 5.5 (0.08) 0.05 6 22 4.9 17 0.23 – (0.31) (0.04) (0.29) (0.01) (0.06) (0.01) (0.06) (0.01) (0.60) (0.09) (0.55) (0.01) 22 February 2011 1 8.4 0.87 7.5 0.10 18 March 2012 5 (1.0) 24 (0.99) (0.19) 2.1 (0.12) (0.93) 22 (1.0) (0.01) 0.09 (0.01) ER, ecosystem respiration; GPP, gross primary production; NEP, net ecosystem productivity. *Metabolic rates could not be estimated for this day due to poor single-station diurnal DO curve fit at one probe location. of previous observations (Mulholland et al., 2001; Bernot et al., 2010). We found that replanted reaches had reduced NEP compared with untreated pasture reaches. Effect of riparian replanting on ecosystem respiration Replanted reaches had greater canopy cover and fine organic matter standing stock. This organic matter resource probably provided an energy source and substratum for microbes and invertebrates. Associations between organic matter supply and increased microbial respiration have been reported in New Zealand forest streams (Young & Huryn, 1999). Large ER rates (up to 32 g O2 m 2 day 1) were observed in a Mediterranean forested stream when autumn led to large amounts of benthic organic matter, but rates were reduced the following year when high flows prevented accumulation (Acu~ na et al., 2004). Accrual of coarse organic matter in replanted reaches of the current study was variable (Fig. 2c). Increased supply of leaf inputs combined with lower average water velocity probably contributed to the accumulation at one replanted reach (Table 1). The variation in our data is not surprising given that the effects of land use on in-stream ER in agricultural catchments compared with forested or low-intensity land-use catchments are still unclear (Young & Huryn, 1999; Young & Collier, 2009; Bernot et al., 2010). ER was similar in streams draining Appalachian agricultural catchments compared with catchments with 50 years of vegetation recovery (McTammany, Benfield & Webster, 2007). Few studies have compared metabolic rate between paired reaches with contrasting riparian condition on one stream. Vegetation removal to restore the natural open-canopy state of US prairie streams reduced ER in some seasons by a magnitude comparable with our results (Riley & Dodds, 2012). ER was c. 1.6 times greater in meadow reaches compared with forested reaches of some US streams (Bott et al., 2006), in contrast to the results from the current study. This may be due to deposition of fine organic matter and increased hyporheic respiration in meadow reaches in the US streams (Bott et al., 2006), whereas benthic FPOM in our streams was greater in replanted reaches. Microbial respiration of DOC contributes much to whole-stream respiration (Wiegner et al., 2005). Although DOC is generally recalcitrant, the addition of labile DOC increased microbial respiration rates and secondary production in a forested headwater (Wilcox et al., 2005). It is unlikely the replanted reaches of the current study are sufficiently extensive to influence DOC quantity detectably. Similar DOC concentrations were observed in the upand downstream reaches of the UR sites, along with FI values indicating dissolved organic matter was from terrestrial vegetation. Labile DOM, leaching from fresh leaf inputs, is readily consumed by the microbial community (Baldwin, 1999). Pulsed inputs of fresh terrestrial organic matter (e.g. from storms), or upstream algal carbon, could increase DOC respiration in replanted reaches. An impor- © 2013 John Wiley & Sons Ltd, Freshwater Biology, doi: 10.1111/fwb.12236 Isolated planting affects ecosystem function Table 3 Linear mixed model results showing mean effect size (SD) of each model parameter (treatment effect, total daily PAR and mean daily water temperature) on the difference in metabolic rates between reaches (downstream – upstream). The estimate for each treatment is the difference between reaches predicted at the overall mean PAR and water temperature NEP OR 0.30 0.32 1.53 0.04 0.41 0.58 0.56 2.41 1.45 1.46 1.40 0.05 0.27 1.38 1.41 1.23 1.4 1.5 6.4 3.6 14 2.2 2.1 22 Treatment (UU) Treatment (RR) Treatment (UR) PAR Temperature Estimate (UU) Estimate (RR) Estimate (UR) 0.18 0.27 0.98 0.00 0.00 0.35 0.25 1.50 1.45 1.46 1.39 0.01 0.05 1.40 1.41 1.22 1.3 1.4 3.4 2.5 1.1 1.5 1.3 9.2 Treatment (UU) Treatment (RR) Treatment (UR) PAR Temperature Estimate (UU) Estimate (RR) Estimate (UR) 0.49 0.61 2.46 0.05 0.38 0.88 0.76 3.82 1.46 1.48 1.49 0.05 0.26 1.41 1.43 1.38 1.8 2.1 16 5.4 14.1 3.2 2.6 54 ER, ecosystem respiration; GPP, gross primary production; NEP, net ecosystem productivity; PAR, photosynthetic active radiation; UU, untreated–untreated; RR, replanted–replanted; UR, untreated– replanted. Important effects and estimates different from zero (OR > 10) are indicated by bold font. tant distinction is whether energy from microbial respiration in the replanted reaches is transferred into secondary production or exported quickly from the reach. This will influence food-web dynamics and affect the effectiveness of replanting on other in-stream functions. The effect of riparian replanting on gross primary production We detected only marginal evidence of lower rates of GPP at replanted reaches compared with untreated reaches at UR streams, despite an increase in canopy cover. Canopy cover, affecting the amount of light reaching the water, often explains a large proportion of variation in GPP (Bunn, Davies & Mosisch, 1999). Canopy cover of the replanted reaches was c. 25% more than upstream untreated reaches and approached the 73% threshold value of canopy cover to yield a P/R value, Δ ER (g O2 m–2 day–1) Treatment (UU) Treatment (RR) Treatment (UR) PAR Temperature Estimate (UU) Estimate (RR) Estimate (UR) GPP SD (a) 4 3 2 1 0 Δ GPP (g O2 m–2 day–1) ER Mean 0 –1 –2 –3 –4 (b) 0 Δ NEP (g O2 m–2 day–1) Parameter Mean (± SE) daily difference in downstream reach Response 5 9 –2 –4 –6 –8 –10 * (c) UU RR Treatment UR Fig. 3 Mean (SE) daily difference in two-station metabolic rates (ecosystem respiration, gross primary production and net ecosystem productivity) between downstream and upstream reaches at the four streams [n = 7, 6, 8 and 5 days for untreated-untreated (UU), replanted–replanted (RR), untreated–replanted (UR) and UR, respectively]. Positive differences indicate greater rates in downstream reaches, while negative values show smaller rates in downstream reaches. Asterisks indicate treatment effects different from zero at UU (white), RR (light grey) and UR (dark grey) from a linear mixed model at an odds ratio >10. indicating a healthy stream in Australia (Bunn et al., 1999). The mean daily proportion of full sunlight reaching the stream surface in all untreated and replanted reaches was 84 and 39%, respectively. This light available below the canopy is much greater than that in many small forested or plantation streams on other continents, which typically have more closed canopies and/ or vegetation with denser foliage (e.g. 1.3% of abovecanopy light; Davies-Colley & Quinn, 1998). Benthic chlorophyll concentrations were similar between reaches at UR sites, suggesting that canopy cover did not affect standing crops of algae or that autotrophs (e.g. diatoms) adapted to lower light conditions are selected for in replanted reaches (Lange et al., 2010). © 2013 John Wiley & Sons Ltd, Freshwater Biology, doi: 10.1111/fwb.12236 10 D. P. Giling et al. Our results differ from studies from the U.S.A. and South America, which reported greater GPP in agricultural than in forested streams (G€ ucker et al., 2009; Bernot et al., 2010). This suggests that other land-use effects, unaffected by small-scale replanting, were important for GPP (see also Riley & Dodds, 2012). A stronger effect of replanting on GPP may not have been detected because considerable channel incision, and hence bank shading, may have limited GPP in all reaches (Young & Huryn, 1999). Erosion has resulted in the deposition of sand and gravel that is unstable and constantly being transported, resulting in low attachment and growth of autotrophs (Atkinson et al., 2008). Total nitrogen and phosphorus exert an important control on GPP (e.g. Frankforter et al., 2010). Nutrient concentrations in our study reaches were not extreme, but generally would indicate a disturbed ecosystem (ANZECC, 2000). Any further increases in nutrient concentration may affect the interaction of GPP and ER between untreated and replanted reaches. This is highly relevant to management because clearance of land for agriculture is associated with increased nutrient concentrations in stream water (Buck, Niyogi & Townsend, 2004). Reducing catchment-scale effects (e.g. erosion and nutrient additions) requires intervention at larger spatial scales than typical replanting projects, which are mostly <1 km in length (Bernhardt et al., 2005). The effect of riparian replanting on NEP and P/R Replanted reaches were more heterotrophic (lower NEP) than untreated pasture reaches, as a result of the shift in both ER and GPP. All study reaches were heterotrophic (mean daily NEP < 0; P/R < 1). Systems with P/R < 1 are not necessarily completely reliant on energy sources from outside the stream because P/R does not account for the source of carbon supporting secondary-consumer respiration (Rosenfeld & Mackay, 1987). The transition to a reliance on energy produced in-stream has been estimated to occur when 0.5 < P/R < 1.0 (Meyer, 1989). All reaches (apart from one untreated reach) had mean daily P/R < 0.5, indicating a reliance on terrestrial or upstream energy sources. This is consistent with results from moderately and some heavily modified pasture streams in the U.S.A. (Hagen et al., 2010). A shift towards reliance on in-stream sources (e.g. Warrenbayne Creek upstream reach; mean P/R = 0.75) in heavily degraded agricultural streams may depend on other factors, such as livestock disturbance or fertiliser inputs (Hagen et al., 2010). The P/R ratio, interpreted appropriately, can indicate the relative importance of organic carbon sources in streams. However, we recommend caution in using P/R to compare paired reaches on different streams because this ratio is sensitive to the magnitude of ER and GPP rates. A small-magnitude difference in opposite directions can have a large effect on the ratio when GPP and ER rates are small, but have a little effect when GPP and ER are large. The difference in NEP between reaches is a more robust measure. A set of criteria to classify stream ecosystem health based on metabolic rates was developed by Young, Matthaei & Townsend (2008). The metabolic rates in most reaches of our study would indicate streams unaffected or mildly affected by land use (Young et al., 2008). Only the downstream reaches of Honeysuckle Creek (low ER rates) and Moonee Creek (high ER rates) could be regarded as being ‘impaired’ (Young et al., 2008). Nevertheless, we saw shifts indicating that replanted reaches had lower NEP rates more typical of unaffected forested streams. Metabolic rates are affected by many proximate factors (e.g. light, nutrients, temperature, organic matter), meaning that responses to interacting stressors can be unclear (e.g. Young & Collier, 2009). This emphasises the importance of using paired reaches to monitor and to assess restoration success. The number of replicate streams places an important limitation on the generality of our results. Longitudinal variation in the direction of response was observed at UU and RR streams, although the effect was small. The response of ER and GPP to vegetation in both UR streams was consistent, suggesting that replanting was responsible. Unidirectional flow means the reaches were not independent, but the vast majority of particulate organic matter transport occurs during storm flow conditions (Wiegner, Tubal & MacKenzie, 2009). Warming can influence NEP (Shurin et al., 2012), but there was no difference in temperature between upstream untreated and downstream replanted reaches at our sites, potentially due to bank shading. Metabolic calculations were made with the change in DO between probes, and the response variable in our statistical analyses was the difference between reaches, which was independent among treatments. What do these results mean for riparian vegetation management? We observed a shift in stream ecosystem processes in response to isolated, reach-scale (i.e. 100s m) patches of riparian vegetation replanting in degraded agricultural catchments. Revegetation can restore the effects of a natural riparian strip, such as shading and organic matter subsidies, even though it is unlikely to influence © 2013 John Wiley & Sons Ltd, Freshwater Biology, doi: 10.1111/fwb.12236 Isolated planting affects ecosystem function non-point effects such as nutrient inflows and altered hydrology (McTammany et al., 2007). These landscape scale land-use effects must be addressed because they may influence metabolic drivers such as sedimentation and DOC availability. Restoring in-stream retentive features (e.g. substratum heterogeneity, debris dams and large wood) in agricultural streams will also be important to ensure that the ecologically beneficial in-stream effects of replanting are realised. Measuring stream ecosystem metabolism integrates organic matter processes, but is less labour intensive than many other biodiversity or functional (e.g. leaf breakdown) measures. We advocate that metabolism should be widely implemented as a functional measure of riparian and stream restoration success (Young et al., 2008; Tank et al., 2010). Recent advances in DO sensors and data processing make longterm installations for monitoring stream health affordable and feasible (Staehr et al., 2010). The influence of stream size on successfully restoring in-stream processes by riparian management is an important issue. Small- to moderate-sized streams, such as those in the current study (second and third order), are expected to respond to riparian restoration more rapidly than large streams and rivers (Craig et al., 2008; Greenwood et al., 2012). We detected changes in metabolic process rates within two decades in streams up to 7.5 m wide. This time frame agrees with simulations of canopy closure following replanting in small channels (DaviesColley et al., 2009). Smaller or shallower channels have a greater capacity to retain coarse and fine organic matter resources (Quinn, Phillips & Parkyn, 2007; Ock & Takemon, 2010). Therefore, we expect metabolic response would take longer and be less pronounced in larger channels. At present, there is little evidence for determining the length of riparian corridor relative to channel width that is needed to influence stream function. The most effective location of restoration for in-stream metabolic process, smaller channels, is also predicted to promote in-stream biodiversity (Death & Collier, 2010). However, we contend that revegetating relatively small and isolated patches may influence stream function, in contrast to recommendations for restoration of biodiversity outcomes (Death & Collier, 2010). The time lag for response of stream ecosystem processes to riparian replanting was shorter than lags observed for recovery of in-stream biodiversity (e.g. Parkyn et al., 2003; Munro et al., 2007). This is important because land required for optimal spatial arrangement of revegetation for biodiversity (i.e. productive, close to existing patches; Thomson et al., 2009) may not always be available for purchase, meaning that restoration may 11 proceed in a ‘piecemeal’ manner. Our results show that, when opportunities arise, land managers should work at a landholder scale to restore riparian vegetation patches and to produce responses in ecosystem processes over short-to-medium terms (decades). Older replantings can be important for biodiversity (Munro et al., 2007), suggesting that revegetated patches may hold biodiversity benefits in future and could form important habitat networks. Monitoring responses to revegetation, in terms of both biodiversity and ecosystem processes, over longer times and in a range of landscape contexts will inform management decisions more effectively and provide insights into the expected timing of responses to ecosystem restoration. Acknowledgments This work was completed with funding from an ARC Linkage Grant (LP0990038) and the Holsworth Wildlife Research Endowment. RT was supported by an ARC Future Fellowship (FT110100957). Advice, field and laboratory assistance was provided by Sam Lake, Shaun Cunningham, Jim Thomson, Samantha Imberger, Adrian Dusting, Laura Caffrey, Glenn Jepson, Scott McDonald, Phil De Zylva, Chris McCormack, Gillian Cromie and Darren Steiert. Alan Hildrew, two anonymous reviewers and members of the Thompson laboratory at Monash University provided valuable comments on earlier versions of this manuscript. References Acu~ na V., Giorgi A., Mu~ noz I., Uehlinger U. & Sabater S. (2004) Flow extremes and benthic organic matter shape the metabolism of a headwater Mediterranean stream. Freshwater Biology, 49, 960–971. ANZECC. (2000) Australian and New Zealand guidelines for fresh and marine water quality. Australian and New Zealand Environment and Conservation Council, Agriculture and Resource Management Council of Australia and New Zealand. APHA. (2005) Standard Methods for the Examination of Water and Wastewater, 21st edn. American Public Health Association, Washington, DC. Asner G.P., Elmore A.J., Olander L.P., Martin R.E. & Harris A.T. (2004) Grazing systems, ecosystem responses, and global change. Annual Review of Environment and Resources, 29, 261–299. Atkinson B.L., Grace M.R., Hart B.T. & Vanderkruk K.E.N. (2008) Sediment instability affects the rate and location of primary production and respiration in a sand-bed stream. Journal of the North American Benthological Society, 27, 581–592. © 2013 John Wiley & Sons Ltd, Freshwater Biology, doi: 10.1111/fwb.12236 12 D. P. Giling et al. Baldwin D.S. (1999) Dissolved organic matter and phosphorus leached from fresh and ‘terrestrially’ aged river red gum leaves: implications for assessing river-floodplain interactions. Freshwater Biology, 41, 675–685. Bernhardt E.S., Palmer M.A., Allan J.D., Alexander G., Barnas K., Brooks S. et al. (2005) Synthesizing U.S. river restoration efforts. Science, 308, 636–637. Bernot M.J., Sobota D.J., Hall R.O., Mulholland P.J., Dodds W.K., Webster J.R. et al. (2010) Inter-regional comparison of land-use effects on stream metabolism. Freshwater Biology, 55, 1874–1890. Bott T.L., Montgomery D.S., Newbold J.D., Arscott D.B., Dow C.L., Aufdenkampe A.K. et al. (2006) Ecosystem metabolism in streams of the Catskill Mountains (Delaware and Hudson River watersheds) and Lower Hudson Valley. Journal of the North American Benthological Society, 25, 1018–1044. Buck O., Niyogi D.K. & Townsend C.R. (2004) Scaledependence of land use effects on water quality of streams in agricultural catchments. Environmental Pollution, 130, 287–299. Bunn S.E. & Davies P.M. (2000) Biological processes in running waters and their implications for the assessment of ecological integrity. Hydrobiologia, 422/423, 61–70. Bunn S.E., Davies P.M. & Mosisch T.D. (1999) Ecosystem measures of river health and their response to riparian and catchment degradation. Freshwater Biology, 41, 333–345. Colloff M.J., Pullen K.R. & Cunningham S.A. (2010) Restoration of an ecosystem function to revegetation communities: the role of invertebrate macropores in enhancing soil water infiltration. Restoration Ecology, 18, 65–72. Craig L.S., Palmer M.A., Richardson D.C., Filoso S., Bernhardt E.S., Bledsoe B.P. et al. (2008) Stream restoration strategies for reducing river nitrogen loads. Frontiers in Ecology and the Environment, 6, 529–538. Davies-Colley R.J., Meleason M.A., Hall R.M.J. & Rutherford J.C. (2009) Modelling the time course of shade, temperature, and wood recovery in streams with riparian forest restoration. New Zealand Journal of Marine and Freshwater Research, 43, 673–688. Davies-Colley R.J. & Quinn J.M. (1998) Stream lighting in five regions of North Island, New Zealand: control by channel size and riparian vegetation. New Zealand Journal of Marine and Freshwater Research, 32, 591–605. Death R.G. & Collier K.J. (2010) Measuring stream macroinvertebrate responses to gradients of vegetation cover: when is enough enough? Freshwater Biology, 55, 1447–1464. Follstad Shah J.J., Dahm C.N., Gloss S.P. & Bernhardt E.S. (2007) River and riparian restoration in the southwest: results of the National River Restoration Science Synthesis Project. Restoration Ecology, 15, 550–562. Frankforter J., Weyers H., Bales J., Moran P. & Calhoun D. (2010) The relative influence of nutrients and habitat on stream metabolism in agricultural streams. Environmental Monitoring and Assessment, 168, 461–479. Gessner M.O. & Chauvet E. (2002) A case for using litter breakdown to assess functional stream integrity. Ecological Applications, 12, 498–510. Greenwood M.J., Harding J.S., Niyogi D.K. & McIntosh A.R. (2012) Improving the effectiveness of riparian management for aquatic invertebrates in a degraded agricultural landscape: stream size and land-use legacies. Journal of Applied Ecology, 49, 213–222. G€ ucker B., Bo€echat I.G. & Giani A. (2009) Impacts of agricultural land use on ecosystem structure and wholestream metabolism of tropical Cerrado streams. Freshwater Biology, 54, 2069–2085. Hagen E.M., McTammany M.E., Webster J.R. & Benfield E.F. (2010) Shifts in allochthonous input and autochthonous production in streams along an agricultural landuse gradient. Hydrobiologia, 655, 61–77. Jeffreys H. (1961) The Theory of Probability, 3rd edn. Clarendon Press, Oxford. Kominoski J.S., Follstad Shah J.J., Canhoto C., Fischer D.G., Giling D.P., Gonzalez E. et al. (in press) Forecasting functional implications of global changes in riparian plant communities. Frontiers in Ecology and the Environment, DOI: 10.1890/120056. Kosinski R.J. (1984) A comparison of the accuracy and precision of several open-water oxygen productivity techniques. Hydrobiologia, 119, 139–148. Lange K., Liess A., Piggott J.J., Townsend C.R. & Matthaei C.D. (2010) Light, nutrients and grazing interact to determine stream diatom community composition and functional group structure. Freshwater Biology, 56, 264–278. Lovett G.M., Cole J.J. & Pace M.L. (2006) Is net ecosystem production equal to ecosystem carbon accumulation? Ecosystems, 9, 152–155. Lunn D.J., Thomas A., Best N. & Spiegelhalter D. (2000) WinBUGS – a Bayesian modelling framework: concepts, structure, and extensibility. Statistics and Computing, 10, 325–337. Marzolf E.R., Mulholland P.J. & Steinman A.D. (1994) Improvements to the diurnal upstream-downstream dissolved oxygen change technique for determining wholestream metabolism in small streams. Canadian Journal of Fisheries and Aquatic Sciences, 51, 1591–1599. McCutchan J.H. & Lewis W.M. (2002) Relative importance of carbon sources for macroinvertebrates in a Rocky Mountain stream. Limnology and Oceanography, 47, 742–752. McKnight D.M., Boyer E.W., Westerhoff P.K., Doran P.T., Kulbe T. & Andersen D.T. (2001) Spectrofluorometric characterization of dissolved organic matter for indication of precursor organic material and aromaticity. Limnology and Oceanography, 46, 38–48. McTammany M.E., Benfield E.F. & Webster J.R. (2007) Recovery of stream ecosystem metabolism from historical agriculture. Journal of the North American Benthological Society, 26, 532–545. © 2013 John Wiley & Sons Ltd, Freshwater Biology, doi: 10.1111/fwb.12236 Isolated planting affects ecosystem function Meyer J.L. (1989) Can P/R ratio be used to assess the food base of stream ecosystems? A comment on Rosenfeld and Mackay (1987). Oikos, 54, 119–121. Mulholland P.J., Fellows C.S., Tank J.L., Grimm N.B., Webster J.R., Hamilton S.K. et al. (2001) Inter-biome comparison of factors controlling stream metabolism. Freshwater Biology, 46, 1503–1517. Munro N.T., Lindenmayer D.B. & Fischer J. (2007) Faunal response to revegetation in agricultural areas of Australia: a review. Ecological Management & Restoration, 8, 199–207. Nusch E.A. (1980) Comparison of methods for chlorophyll €r Hydrobiologie– and phaeopigment determination. Archiv fu Beiheft Ergebnisse der Limnologie, 14, 14–36. Ock G. & Takemon Y. (2010) Estimation of transport distance of fine particulate organic matter in relation to channel morphology in tailwaters of the Lake Biwa and reservoir dams. Landscape and Ecological Engineering, 6, 161–169. Odum H.T. (1956) Primary production in flowing waters. Limnology and Oceanography, 1, 102–117. Parkyn S.M., Davies-Colley R.J., Halliday N.J., Costley K.J. & Croker G.F. (2003) Planted riparian buffer zones in New Zealand: do they live up to expectations? Restoration Ecology, 11, 436–447. Poiani K.A., Richterm B.D., Anderson M.G. & Richter H.E. (2000) Biodiversity conservation at multiple scales: functional sites, landscapes, and networks. Bioscience and Microflora, 50, 133–146. Quinn J.M., Phillips N.R. & Parkyn S.M. (2007) Factors influencing retention of coarse particulate organic matter in streams. Earth Surface Processes and Landforms, 32, 1186– 1203. Riley A.J. & Dodds W.K. (2012) The expansion of woody riparian vegetation, and subsequent stream restoration, influences the metabolism of prairie streams. Freshwater Biology, 57, 1138–1150. Roberts B., Mulholland P. & Hill W. (2007) Multiple scales of temporal variability in ecosystem metabolism rates: results from 2 years of continuous monitoring in a forested headwater stream. Ecosystems, 10, 588–606. Rosenfeld J.S. & Mackay R.J. (1987) Assessing the food base of stream ecosystems: alternatives to the P/R ratio. Oikos, 50, 141–147. Shurin J.B., Clasen J.L., Greig H.S., Kratina P. & Thompson P.L. (2012) Warming shifts top-down and bottom-up control of pond food web structure and function. Philosophical Transactions of the Royal Society B: Biological Sciences, 367, 3008–3017. Staehr P., Testa J., Kemp W., Cole J., Sand-Jensen K. & Smith S. (2010) The metabolism of aquatic ecosystems: history, applications, and future challenges. Aquatic Sciences – Research Across Boundaries, 74, 15–29. Stoate C., Baldi A., Beja P., Boatman N.D., Herzon I., van Doorn A. et al. (2009) Ecological impacts of early 21st century agricultural change in Europe – a review. Journal of Environmental Management, 91, 22–46. Tank J.L., Rosi-Marshall E.J., Griffiths N.A., Entrekin S.A. & Stephen M.L. (2010) A review of allochthonous organic matter dynamics and metabolism in streams. Journal of the North American Benthological Society, 29, 118–146. Thompson R.M. & Townsend C.R. (2004) Land-use influences on New Zealand stream communities: effects on species composition, functional organisation, and foodweb structure. New Zealand Journal of Marine and Freshwater Research, 38, 595–608. Thomson J.R., Moilanen A.J., Vesk P.A., Bennett A.F. & Mac Nally R. (2009) Where and when to revegetate: a quantitative method for scheduling landscape reconstruction. Ecological Applications, 19, 817–828. U.S. Census Bureau. (2004) Global Population Profile: 2002, Int. Popul. Rep. WP/02. Washington, DC. Wiegner T.N., Kaplan L.A., Newbold J.D. & Ostrom P.H. (2005) Contribution of dissolved organic C to stream metabolism: a mesocosm study using 13C-enriched treetissue leachate. Journal of the North American Benthological Society, 24, 48–67. Wiegner T.N., Tubal R.L. & MacKenzie R.A. (2009) Bioavailability and export of dissolved organic matter from a tropical river during base- and stormflow conditions. Limnology and Oceanography, 54, 1233–1242. Wilcox H.S., Wallace J.B., Meyer J.L. & Benstead J.P. (2005) Effects of labile carbon addition on a headwater stream food web. Limnology and Oceanography, 50, 1300–1312. Young R.G. & Collier K.J. (2009) Contrasting responses to catchment modification among a range of functional and structural indicators of river ecosystem health. Freshwater Biology, 54, 2155–2170. Young R.G. & Huryn A.D. (1999) Effects of land use on stream metabolism and organic matter turnover. Ecological Applications, 9, 1359–1376. Young R.G., Matthaei C.D. & Townsend C.R. (2008) Organic matter breakdown and ecosystem metabolism: functional indicators for assessing river ecosystem health. Journal of the North American Benthological Society, 27, 605–625. Supporting Information Additional Supporting Information may be found in the online version of this article: Data S1. Model description. (Manuscript accepted 4 August 2013) © 2013 John Wiley & Sons Ltd, Freshwater Biology, doi: 10.1111/fwb.12236 View publication stats 13