Oecologia DOI 10.1007/s00442-015-3279-5

COMMUNITY ECOLOGY - ORIGINAL RESEARCH

Autumn leaf subsidies influence spring dynamics of freshwater plankton communities Samuel B. Fey · Andrew N. Mertens · Kathryn L. Cottingham 

Received: 13 August 2014 / Accepted: 18 February 2015 © Springer-Verlag Berlin Heidelberg 2015

Abstract  While ecologists primarily focus on the immediate impact of ecological subsidies, understanding the importance of ecological subsidies requires quantifying the long-term temporal dynamics of subsidies on recipient ecosystems. Deciduous leaf litter transferred from terrestrial to aquatic ecosystems exerts both immediate and lasting effects on stream food webs. Recently, deciduous leaf additions have also been shown to be important subsidies for planktonic food webs in ponds during autumn; however, the inter-seasonal effects of autumn leaf subsidies on planktonic food webs have not been studied. We hypothesized that autumn leaf drop will affect the spring dynamics of freshwater pond food webs by altering the availability of resources, water transparency, and the metabolic state of ponds. We created leaf-added and no-leaf-added field mesocosms in autumn 2012, allowed mesocosms to ice-over for the winter, and began sampling the physical, chemical, and biological properties of mesocosms immediately following ice-off in spring 2013. At ice-off, leaf additions reduced dissolved oxygen, elevated total phosphorus concentrations and dissolved materials, and did not alter temperature or total nitrogen. These initial abiotic effects

contributed to higher bacterial densities and lower chlorophyll concentrations, but by the end of spring, the abiotic environment, chlorophyll and bacterial densities converged. By contrast, zooplankton densities diverged between treatments during the spring, with leaf additions stimulating copepods but inhibiting cladocerans. We hypothesized that these differences between zooplankton orders resulted from resource shifts following leaf additions. These results suggest that leaf subsidies can alter both the short- and longterm dynamics of planktonic food webs, and highlight the importance of fully understanding how ecological subsidies are integrated into recipient food webs. Keywords  Terrestrial-aquatic linkages · Food webs · Phenology · Ponds · Zooplankton

“The rain falling on the freshly dried herbs and leaves, and filling the pools and ditches into which they have dropped thus clean and rigid, will soon convert them into tea,— green, black, brown, and yellow teas, of all degrees of strength, enough to set all Nature a-gossiping.” H. D. Thoreau, "Autumnal Tints,” 1862.

Communicated by Robert O. Hall. S. B. Fey and A. N. Mertens have contributed equally to this manuscript. S. B. Fey · A. N. Mertens · K. L. Cottingham  Department of Biological Sciences, Dartmouth College, Hanover, NH, USA Present Address: S. B. Fey (*)  Department of Ecology and Evolutionary Biology, Yale University, New Haven, CT 06520, USA e-mail: [email protected]

Introduction Ecological subsidies, the energy and materials that move across ecosystem boundaries, influence the structure and composition of food webs (Polis et al. 1997; Takimoto et al. 2002; Gratton et al. 2008). The effect of such subsidies on the recipient community depends on many factors, including the characteristics (Cole et al. 2006) and quantity of materials transferred (Marcarelli et al. 2011), the rate

13

Oecologia

of subsidization (Takimoto et al. 2009), and the behavioral and life history characteristics of the recipient community (Baxter et al. 2005). Tremendous variation exists regarding the time period over which a subsidy can affect the recipient food web. For example, adult insects emerging from temperate aquatic ecosystems into terrestrial ecosystems generally integrate rapidly into the terrestrial food web (Nakano and Murakami 2001; Sabo and Power 2002). By contrast, coarse woody debris that enters aquatic ecosystems from surrounding forests may decay over decades or even centuries (Guyette and Cole 1999). Similarly, the integration of large whale carcasses (which can fall from pelagic habitats following death) into benthic ecosystems first by mobile scavengers, then polychaetes and crustaceans, and finally microbes, can take place over decades (Smith and Baco 2003; Higgs et al. 2014). As such, accurately quantifying the effect of a cross-ecosystem subsidy may require quantifying the recipient ecosystem’s response both immediately and over extended time periods. Leaves from deciduous trees represent an important subsidy for aquatic ecosystems that alters food webs on both short and long time scales (Anderson and Sedell 1979; Earl et al. 2014). Allochthonous inputs to streams from surrounding forests are well appreciated and typically exceed autochthonous inputs in certain ecosystems (Webster and Meyer 1997). By providing both limiting resources and microhabitat (Richardson 1992), leaves aid in the immediate production of bacteria, fungi, and detritivorous insects in streams (Hieber and Gessner 2002). The use of leaf packs containing terrestrial leaf litter also provides strong evidence that leaf subsidies are integrated into recipient ecosystems over long time periods, as shown by how the remaining leaf material continues to decrease in mass following multiple seasons of being utilized by stream biota (Benfield et al. 2001; Swan and Palmer 2004). Moreover, excluding terrestrial leaf litter inputs significantly reduces benthic invertebrate biomass, including shredders, gatherers, and even predators on interseasonal timescales (Wallace et al. 1999). In temperate pond ecosystems, autumn leaf drop can contribute substantially to annual carbon and nutrient budgets (France and Peters 1995; Hongve 1999; Pope et al. 1999), particularly in ecosystems unable to produce large amounts of autochthonous carbon (Rubbo et al. 2006). Furthermore, while there is a growing appreciation for the importance of leaves as a subsidy that immediately impacts producer and consumer components of pond food webs (Oertli 1993; Rubbo et al. 2008; Rubbo and Kiesecker 2004; Cottingham and Narayan 2013; Earl et al. 2014), little is known about the long-term effects of leaf subsidies for such food webs. Our current understanding of this effect comes primarily from two whole-ecosystem experiments that investigated the effect of leaf litter on small ponds by removing or restricting the fall of leaf litter in autumn, and measuring the subsequent

13

Fig. 1  Leaves and leaf leachates impact pond food webs by resourcemediated effects (solid black arrows), transparency-mediated effects (dashed grey arrows), and oxygen-mediated effects (solid grey arrows)

response in the spring. Here, Rubbo et al. (2006) showed that leaf removals decreased ecosystem respiration and reduced inter-seasonal secondary production. However, high between-pond variability produced contrasting results for a similar study investigating the effect of leaf litter on benthic invertebrates (Batzer and Palik 2007). To our knowledge, no study has addressed the long-term effect of leaves on organisms other than benthic invertebrates, and especially not on both the autotroph- and detritus-based food webs. There is reason to believe that autumn leaf subsidies could exert effects on planktonic food webs that extend beyond the autumn months, particularly in temperate ecosystems where reduced light availability and cold temperatures in winter are likely less conducive for planktonic food webs to assimilate the carbon and nutrients provided by autumn leaf subsidies (Twiss et al. 2012; Bertilsson et al. 2013). We identify three pathways through which leaves may affect pond food webs (Fig. 1). First, resource-mediated effects result from the consumption of leaf material, the uptake of nutrients and carbon from leaf leachates, or the consumption of organisms that uptake these materials. Second, transparency-mediated effects result from changes in water transparency that regulates the penetration of incoming solar radiation. Third, metabolism-mediated effects result from changes in the balance of respiration versus photosynthesis in response to leaf additions. The combined changes to the biotic and abiotic environment would have cascading but differential effects on zooplankton orders. Assuming these effects persist into the spring, we predict that cladocerans will benefit relative to copepods due to decreased ultraviolet radiation exposure accompanying leaf additions (Leech and Williamson 2000), cladocerans will be negatively impacted to a greater extent than copepods due to the decreases in oxygen concentrations (Vanderploeg et al. 2009) that accompany leaf additions (Cottingham and Narayan 2013), and the faster numeric responses of cladocerans would allow them to benefit from any increases in phytoplankton stimulated by leaf additions.

Oecologia

Therefore, we hypothesized that autumn leaf drop affects the spring dynamics of pond food webs by altering the availability of resources, the water transparency, and the metabolic state of ponds. We conducted a 10-month mesocosm experiment using an experimental array previously used to understand how variation in leaf chemistry driven by increased soil temperature can alter the quality of leaf subsidies for planktonic food webs (Fey et al. 2015). Here, we test two specific predictions: autumn leaf subsidies affect the initial state of planktonic food webs immediately following ice-off, and have a lasting effect on planktonic food webs beyond ice-off.

Materials and methods Our field experiment involved several stages: 1. Establishing mesocosms and imposing the control (5 g leaves added to mesocosms) and leaf-addition (50 g leaves added to mesocosms) experimental treatments. 2. Following responses to treatments prior to ice-on (Fey et al. 2015). 3. Allowing mesocosms to freeze over during the winter months. 4. Observing the treatment responses immediately following ice-out and throughout the spring. There were five replicates of each treatment, for a total of ten mesocosms. The primary response variables were phytoplankton chlorophyll-a, total nitrogen and total phosphorus concentrations, bacterial and zooplankton densities, and zooplankton community composition. We also characterized the abiotic environment by measuring dissolved oxygen, temperature, and the absorbance due to dissolved substances at 254 nm—a proxy for water clarity and dissolved organic carbon (Brandstetter et al. 1996). The leaf loading rate used (~150 g leaves/m2 pond surface area) falls within the reported rates of areal loading rates observed in the near-shore areas of lakes (Hongve 1999) and previous leaf-addition mesocosm experiments (Rubbo and Kiesecker 2004; Cottingham and Narayan 2013). These leaf additions correspond to a loading rate of ~23 g leaves per meter “shoreline” (as measured by mesocosm circumference), which falls below the 32–720 g/m reported annual shoreline rates (France and Peters 1995). Establishing mesocosm communities We constructed mesocosms from 167-L polyethylene cylindrical refuse bins (0.61-m diameter by 0.80-m depth, 1.95 surface to volume ratio) buried in the ground at the Dartmouth Organic Farm, Hanover, New Hampshire.

Mesocosms were filled with 120 L of groundwater on 5 September 2012 and covered with 1-mm fiberglass mesh to prevent colonization by non-target organisms. The following day, we added homogenized packets containing 5 g of the collected leaf subsidies to all mesocosms to provide a source of organic carbon and nutrients for developing biological communities. On 6–7 September 2012, we added 5 L of water from each of six small lakes and ponds within a 20-km radius of Hanover (30 L total, final mesocosm volume ~150 L) to seed mesocosms with natural phytoplankton and bacterial communities (Table S1). From 9 to 11 September, we added zooplankton from both the littoral (near shore) and pelagic (open water) zones of five of the six ponds. We collected zooplankton using a 12-L Schindler-Patalas trap; each mesocosm received zooplankton filtered from 144 L of pond water. Prior to imposing leaf-addition treatments, there were no differences between mesocosm treatments in any of our response variables (Fey et al. 2015). We did not add benthic invertebrates to our mesocosms in order to focus on the response of the pelagic food web. Generating experimental treatments We collected leaves from the Harvard Forest in central Massachusetts, USA (42°28′N, 72°10′W), rather than from the shoreline of the ponds used to establish mesocosms, to facilitate a prior experiment (Fey et al. 2015). On 28 October 2010 and 20 October 2011, we collected freshly fallen leaves from red maple trees (Acer rubrum), which are common around ponds across eastern North America (Rubbo and Kiesecker 2004). All leaves were collected after abscission but before a rain event. Leaves were dried at room temperature, and then stored in closed paper bags until the experiment so as not to alter leaf chemistry (Cotrufo and Ineson 1996). Leaf subsidies were added to mesocosms on 2 October 2012. We added 45 g of dried leaf litter (10 g from 2010 and 35 g from 2011) to each mesocosm in the leaf-addition treatment. Field sampling We monitored mesocosms through the fall season and noted that mesocosms were fully iced-over the week of 13 November (Fey et al. 2015). Ice-off occurred in all mesocosms during the last week of April 2013. While complete surface ice coverage of the mesocosms persisted during the winter, we did not measure the thickness of the ice to avoid disrupting under-ice dynamics. As the ice and snow melted, we removed additional water to ensure that the mesocosm volume remained at 150 L and that water did not make contact with the fiberglass screens enclosing mesocosms.

13

Oecologia

Beginning on 1 May 2013, we sampled mesocosms weekly between 0800 and 1200 hours. We recorded water temperature, dissolved oxygen, using a YSI model 556 MPS meter (YSI, Yellow Springs, OH). We then homogenized the mesocosms through stirring, removed a 5-L vertical cross-section of the water column using a vertical water sampler, and filled 3 1-L bottles for nutrient, chlorophyll, and bacterial analysis, respectively. All samples were stored in coolers containing ice packs until returning to the laboratory. We collected an additional 5 L of water to sample for zooplankton. We filtered the water through 53-μm mesh, washed the zooplankton off the mesh into 50-mL centrifuge tubes, preserved the zooplankton with 70 % ethanol, and returned the filtered water to the mesocosm. Mesocosm sampling ended at the end of spring, 19 June, when treatment and control mesocosms had converged in most physical and chemical properties. Laboratory analyses Nutrients and chlorophyll-a were analyzed followed previously described procedures (Fey and Cottingham 2012). Total nutrient samples were frozen until digestion and spectrophotometric analysis on a Perkin-Elmer Lambda EZ201 spectrophotometer (PerkinElmer, Waltham, MA). For total nitrogen, we used a basic persulfate digestion and measured nitrate using the second-derivative method (Crumpton et al. 1992). For total phosphorus, we used a persulfate digestion and the molybdate colorimetric reaction. Samples for dissolved materials were pre-filtered using PURADISC 25 GF/F disposable filters (Whatman, Maidstone, Kent, UK), and were measured spectrophotometrically by absorbance in a 1-cm quartz cuvette at 254 nm (Brandstetter et al. 1996). Finally, to measure chlorophyll-a concentration, we vacuum-filtered 100 mL of water onto 47-mm Whatman GF/C filters; froze the filters, extracted the chlorophyll in methanol, and measured concentrations using the Welschmeyer method (Welschmeyer 1994). Zooplankton were enumerated at 10× magnification on a Leica MZ-12 dissecting microscope (Leica Microsystems, Bannockburn, IL). Cladocerans were identified to genus; adult copepods were identified to order (Calanoida versus Cyclopoida), while copepodites were enumerated as a bulk category. We did not detect any rotifers in mesocosms, similar to previous mesocosm experiments (Fey and Cottingham 2012; Cottingham and Narayan 2013). Bacteria were enumerated using a modified 4′,6-diamidino-2-phenylindole (DAPI) staining method (Porter and Feig 1980). Upon returning from the field, whole water samples were fixed immediately with glutaraldehyde (3 % final concentration; Acros Organics, Geel, Belgium). Prior to enumeration, samples were filtered at low pressure (<0.3 bar) onto 0.2-µm Isopore black polycarbonate

13

filters (Millipore, Billerica, MA), placed on glass slides, and stored at −20  °C. Bacteria were then visualized using a Nikon Eclipse 90i microscope (Nikon Instruments, Melville, NY) with an ET DAPI filter [excitation filter (350/50), dichroic mirror (400 LP), emission filter (460/50); Chroma Technology, Bellows Falls, VT]. Nine 89.7 × 67-µm quadrants of each membrane were photographed with a Photometrics CoolSnap HQ2 integrated camera (Photometrics, Tucson, AZ) and bacteria were enumerated in each image using ImageJ 1.46r. Bacterial densities in each sample were determined from the mean density across the nine quadrants. Statistical analyses To address our prediction that autumn leaf subsidies affect the initial state of planktonic food webs immediately following ice-off, we used unequal variance two-sample t-tests to determine whether there were treatment differences immediately following ice-out (1 May) and on the last sampling date (19 June) for all response variables. We transformed all count data (zooplankton, log10 + 0.2 individuals per liter; bacteria, log10 cells per milliliter) data to reduce variance inequality. Secondly, to test the prediction that autumn leaf subsidies have a lasting effect on planktonic food webs throughout the spring, we used mixed-models repeated-measures ANOVA (RM-ANOVA; R Statistics Environment, nlme package) to test for significant time × treatment interaction effects on each response variable. We selected the covariance structure among sampling dates (e.g., how observations made at short time intervals are correlated relative to more distant observations) using the Bayesian information criterion (Quinn and Keough 2002), choosing between compound symmetry, unstructured, autoregressive heterogeneous variances, and autoregressive covariance structures. All analyses were conducted in R 3.1.0 (R Foundation for Statistical Computing, Vienna) and treatment effects were considered significant at α = 0.05.

Results Leaf subsidies’ initial impact following ice‑off Autumn leaf subsidies yielded differences between experimental treatments immediately following ice-off, which typically resembled autumn conditions prior to ice-on (Fig. 2). The absorbance of dissolved substances measured at 254 nm (1/m), which relates to dissolved organic carbon concentration (Brandstetter et al. 1996), was 20 times lower in treatments receiving leaves immediately after ice-out

Oecologia

Fig. 2a–c  Abiotic responses to leaf additions in experimental mesocosms. Dissolved O2 is measured in mg/L. Black squares are data from leaf-addition mesocosms and white circles are from control mesocosms. Left panels are data from 13 November 2012, the last sample date before ice-over (Fey et al. 2015). Data are treatment mean ± SE (n = 5 replicates). A254 Absorbance from dissolved substances at 254 nm (1/m)

(Fig. 2a; t8.76 = 3.00, P = 0.003). Dissolved oxygen concentrations were initially 40 % lower in mesocosms receiving leaf additions (Fig. 2b; t6.30  =  −24.62, P < 0.001). Temperature did not differ between experimental treatments (Fig. 2c; t6.63 = 0.37, P > 0.5). Leaf additions resulted in elevated total phosphorus but not total nitrogen at ice-off (Fig. 3). The difference in nitrogen concentrations between treatments converged over winter such that nitrogen concentrations did not differ immediately following ice-off (Fig. 3a; ice-out t5.45 = 0.50, P > 0.6). The effect of leaf additions on total phosphorus persisted over winter, such that total phosphorus in the leaf-addition treatment concentrations remained five times higher immediately following ice-off (Fig. 3b; ice-out t4.06 = 5.57, P = 0.005). Leaf additions initially altered chlorophyll-a concentrations and bacterial densities following ice-off. Leaf-addition treatments initially had reduced chlorophyll-a concentrations by 50 % (Fig. 4a; t5.55  =  −3.73, P  = 0.011). By contrast, bacterial densities were two-fold higher in

Fig. 3a, b  Total phosphorus and nitrogen concentrations through the spring for leaf-addition and control mesocosms. Black squares are data from leaf-addition mesocosms and white circles are data from control mesocosms. Left panels are data from 13 November 2012, the last sample date before ice-over. Data are treatment mean ± SE (n = 5 replicates)

mesocosms receiving leaves immediately following ice-off (Fig. 4b; t7.57 = 3.20, P = 0.014). Leaf additions initially decreased cladoceran densities to near-zero densities relative to control treatments, although cladoceran densities were initially low in both treatments (Fig. 5a, c, e, g). Consequently, total cladoceran, Chydorus, and Daphnia densities were higher in control treatments immediately following ice-off (t3.15  =  −4.97, P  = 0.014; t3.83  =  −4.55, P  = 0.012; and t6.97  =  −1.09, P > 0.3, respectively), while Bosmina densities did not differ between treatments (t6.96  =  −10.05, P < 0.001). In contrast to cladocerans, leaf additions had a minimal effect on copepod densities immediately following ice-off, consistent with the lack of response to leaves prior to ice-on (Fig.  5b, d, f, h). There was no treatment difference for total copepods, juvenile copepods, calanoids, or cyclopoids immediately following ice-off (all P > 0.3). Autumn leaf subsidies impact throughout the spring By late spring, most abiotic variables had converged or were converging between treatments. The absorbance of dissolved substances measured at 254 nm (1/m) decreased throughout the spring in the leaf-addition treatment,

13

Oecologia

Fig. 4a, b  Basal resource response to leaf addition, as measured by chlorophyll-a concentration and bacteria densities. Black squares are data from leaf-addition mesocosms and white circles are data from control mesocosms. Left panels are data from 13 November 2012, the last sample date before ice-over (Fey et al. 2015). Data are treatment mean ± SE (n = 5 replicates)

while the control treatment did not change (Fig. 2a; RMANOVA time × treatment F7,59  = 10.89, P < 0.001). As such, absorbance remained fivefold higher in mesocosms receiving leaves at the end of spring (t4.57 = 4.32, P  = 0.009). Dissolved oxygen concentrations increased during the spring in leaf-addition treatments, while decreasing in control treatments (Fig. 2b; RM-ANOVA time  × treatment F7,64  = 28.82, P < 0.001), such that no difference existed by the end of spring (t6.61  = 0.41, P > 0.5). There was no treatment effect on temperature throughout the spring (Fig. 2c; RM-ANOVA treatment F1,64  = 0.156, P > 0.6; end of spring, t6.61  =  −1.24, P > 0.2). Despite the initial differences in phosphorus to leaf additions at ice-off, the chemical environments between treatments converged by the end of spring. Nitrogen concentrations increased similarly in both treatments (Fig. 3a; RM-ANOVA time F7,64  = 9.55, P < 0.001) and again did not differ at the end of the experiment (t6.59  = 1.84, P > 0.1). Phosphorus concentrations decreased in leafaddition treatments, while increasing in control treatments (Fig.  2b; RM-ANOVA time × treatment F7,64  = 11.32, P < 0.001), such that no treatment effects existed by the end of spring (t7.20 = 0.33, P > 0.7).

13

While chlorophyll-a concentrations differed immediately following ice-off, concentrations increased faster in leaf-addition treatments (Fig. 4a; RM-ANOVA time × treatment F7,64 = 4.85, P < 0.001) and did not differ by the end of spring (t6.57 = 1.13, P > 0.2). Bacterial densities, however, followed the opposite trajectory. Bacteria densities were generally higher in leaf-addition mesocosms (Fig. 4b; RM-ANOVA treatment F1,64  = 10.283, P  = 0.003) and increased more rapidly in the control treatment throughout the spring (Fig. 4b; RM-ANOVA time  × treatment F4,64  = 5.26, P  = 0.002), reaching densities 50 % higher in control than leaf-addition mesocosms by the end of spring (Fig. 4b; t6.98  =  −2.44, P = 0.045). In contrast to the physical and chemical environment, zooplankton populations generally continued to diverge between treatments as the spring progressed. Leaf additions continued to suppress cladoceran densities compared to control treatments. Total cladoceran and Chydorus population densities were higher in control mesocosms throughout the spring (Fig. 5a, e; RM-ANOVA treatment F1,64  = 26.81, P < 0.001 and F1,64  = 56.61, P < 0.001, respectively), and were 70 and 15 individuals/L higher at the end of the experiment, respectively (t7.25  =  −2.75, P  = 0.027 and t7.91  =  −5.21, P < 0.001, respectively). Bosmina densities, which were very low in all mesocosms at the start of spring sampling, rapidly increased in control mesocosms while remaining low in leaf-addition mesocosms (Fig. 5c; RM-ANOVA time × treatment F7,64  = 3.825, P  = 0.002). Overall, Daphnia densities were elevated in no-leaf treatments throughout the spring (Fig. 5g, RM-ANOVA treatment F1,64 = 13.35, P < 0.001), although no difference existed by the end of the spring (Fig. 5g t5.62 = −1.88, P > 0.1). While leaf additions suppressed cladoceran population growth, they generally increased copepod growth (Fig. 5). Total copepod, juvenile copepod, and cyclopoids increased more rapidly in leaf-addition treatments (Fig. 5b, d, h; RM-ANOVA time × treatment F7,64  = 6.62, P < 0.001; F7,64  = 2.77, P  = 0.014; and F7,64  = 6.37, P < 0.001, respectively), and tended to be higher (2, 5 and 2× higher, respectively) on the last day of sampling, (t7.94  = 2.27, P = 0.053; t7.97 = 2.35, P = 0.047; t7.01 = 2.45, P = 0.044, respectively). Calanoids were suppressed by leaf additions, with leaf additions leading to reduced calanoids densities throughout the spring (Fig. 5f; RM-ANOVA treatment F1,64 = 6.316, P = 0.016). Discussion Our experimental results supported the hypothesis that autumn leaf drop can affect spring pond dynamics by

Oecologia Fig. 5  Cladoceran (a, c, e, g) and copepod (b, d, f, h) responses to leaf additions. Black squares are data from leaf-addition mesocosms and white circles are data from control mesocosms. The left panels are data from 13 November 2012, the last sample date before ice-over (Fey et al. 2015). Data are treatment mean ± SE (n = 5 replicates)

creating persistent differences in the availability of resources, oxygen concentrations, and water transparency (Fig. 1). At ice-off, leaf additions elevated total phosphorus, yet the abundance of dissolved substances in mesocosms receiving leaves limited light availability and depressed chlorophyll concentrations, thereby acting as a stress (Odum et al. 1979) on producers. While the physical and chemical conditions converged between treatments as the spring progressed, bacteria and zooplankton continued to diverge through time between leaf addition treatments, with leaf additions stimulating copepods but depressing cladoceran zooplankton and bacteria. We now hypothesize this difference in zooplankton performance is due to variation in the ability of zooplankton orders to selectively use food resources in the presence of high quantities of dissolved and particulate substances (Wilkinson et al. 2013; Berggren et al. 2014). As such, our results indicate that leaf subsidies affect the initial state of pond food webs at ice-out and may have a lasting effect throughout the spring season.

Inter‑seasonal leaf effects on the abiotic environment and basal food web components Initial conditions immediately following ice-off strongly resembled the mesocosm food web structure prior to iceon. Indeed, the components that were at higher abundance in mesocosms receiving leaves prior to ice-on in the fall remained more abundant in the spring. These initial spring conditions of reduced chlorophyll but increased bacterial densities, combined with elevated phosphorus and dissolved materials, were consistent with previous mesocosm studies investigating the autumnal effects of leaf litter on aquatic food webs (Pope et al. 1999; Rubbo and Kiesecker 2004; Cottingham and Narayan 2013). Our research extends France and Peters (1995) study of how ponds respond with increased respiration immediately following autumn leaf drop, and suggests that in temperate ecosystems, autumn leaf additions may continue to affect pond metabolism during spring months. Yet, it will be of further

13

Oecologia

interest to measure the inter-seasonal dynamics of metabolism in response to leaf addition. As spring progressed, the leaf additions appeared to shift from acting as a stress to a subsidy (Odum et al. 1979) for the autotroph-based food web. As the physical and chemical environments between leaf and no-leaf treatments converged, total chlorophyll and total phosphorus converged between experimental treatments, providing evidence that phytoplankton began capitalizing on the nutrients provided by leaves. The importance of light limitation as a limiting factor for the primary production for certain aquatic ecosystems (Karlsson et al. 2009) and the importance of nitrogen and phosphorus in fueling primary production (Elser et al. 2007) likely contributed to this increase. Concurrently, the convergence of bacterial densities between experimental treatments likely resulted from the growing contribution of decaying autochthonous phytoplankton and the extracellular release of dissolved organic matter from phytoplankton for bacterial growth in mesocosms not receiving autumn leaf additions (Ueveges et al. 2012). Inter‑seasonal effects of leaf subsidies on zooplankton That leaf additions initially inhibited cladocerans but had no effect on copepod densities was consistent with how zooplankton orders immediately responded to autumn leaf additions. Cladoceran zooplankton have been shown to be more sensitive to the effects of leaf additions relative to copepods over short time scales in other mesocosm experiments (Cottingham and Narayan 2013). This pattern may be attributed to an accelerated numeric response of cladoceran zooplankton, which have faster generation times relative to copepods, during the autumn months (Williamson and Reid 2009); however, the differences between treatments observed in the end of the fall persisted throughout the winter months. The presence of zooplankton immediately following ice-out mimics natural ponds where zooplankton are able to persist through winter under the ice despite low temperatures, limited light, and food availability (Karlsson et al. 2009). We do not believe that the direct effects of both oxygen and light limitation (the absorbance of dissolved materials at 254 nm) account for the observed differences in zooplankton between treatments, despite generating treatment differences. Cladocerans are considered to be more susceptible to the effects of low oxygen concentrations (Vanderploeg et al. 2009) and as such might be expected to perform worse in the leaf-addition treatment due to the effects of oxygen alone. However, the relatively low starting concentrations of dissolved oxygen (~8 mg/L) in mesocosms with added leaves are still four times higher than oxygen concentrations cladocerans typically avoid (Vanderploeg et al. 2009). Thus, oxygen-mediated effects (Fig. 1) present

13

an unlikely explanation for the varied responses of cladocerans versus copepods. Likewise, it is unlikely that the decreases in light availability in the leaf-addition treatment resulted in the elevated copepod densities because copepods are more tolerant of the harmful effects of ultraviolet radiation than cladocerans (Leech and Williamson 2000), and thus copepods might be expected to benefit relative to cladocerans in mesocosms not receiving leaf additions. Instead, we observed the opposite pattern expected from transparency-mediated effects (Fig. 1), with leaf additions depressing cladocerans but stimulating copepods throughout the spring months. Importantly, the trajectories zooplankton followed throughout the spring in response to leaves differed from the dynamics typically exhibited by zooplankton in autumn (Fey et al. 2015). As the spring progressed following iceoff, total cladoceran densities and total copepod densities exhibited compensatory dynamics, such that leaf additions increased copepod densities, but reduced cladoceran densities, relative to control mesocosms. This pattern was in contrast to the response of zooplankton orders during the autumn in a separate leaf mesocosm experiment (Cottingham and Narayan 2013), where leaf additions stimulated cladoceran zooplankton. The delayed positive response of copepods to leaf additions, which did not materialize until well into the spring, indicates the importance of observing long-term dynamics of the response to leaf additions. One plausible explanation for the general divergence of copepod and cladoceran densities in response to leaf additions lies in their methods of feeding. Copepods are generally more selective feeders than the primarily filter-feeding omnivorous cladocerans found in our mesocosms (Wilkinson et al. 2013; Berggren et al. 2014). Additionally, copepods are thought to be more predatory, and may therefore eat larger components of the microbial loop such as ciliates and heterotrophic nano-flagellates (Williamson and Reid 2009). As such, they may have benefited from hunting the abundance of phytoplankton and microbial resources in treatment mesocosms, while cladocerans may have accumulated small leaf particulates in their guts during their normal filter feeding (Lennon et al. 2013). Additionally, the observations that terrestrial organic carbon concentrations negatively affect cladocerans, but impact copepods to a lesser extent in natural lakes (Kelly et al. 2014), and that copepods use terrestrial organic carbon to a lesser extent than cladocerans (Kelly et al. 2014) support the finding that leaf additions may act as more of a stress than a subsidy for cladocerans. Subsidy utilization by pond planktonic food webs While our experiment demonstrates the ability of leaves to affect planktonic food webs on inter-seasonal time scales,

Oecologia

the mechanisms underlying the long-term effect of autumn leaf additions on the planktonic food webs remain elusive. Many ecological subsidies with documented long-term effects on the recipient ecosystem are often recalcitrant (e.g., coarse woody debris in aquatic ecosystems), slowly and continuously integrating into recipient ecosystems over long time periods. By contrast, nutrients and carbon leach from leaves within a few days of entering aquatic ecosystems (Rubbo et al. 2006; Cottingham and Narayan 2013). The persistent effect of leaf leachates in our mesocosms may result instead from lack of subsidy use by the recipient community due to biological limitations (Twiss et al. 2012; Bertilsson et al. 2013). The very conditions that initiate autumn leaves drop in temperate regions—cold temperatures and reduced solar radiation—are also associated with reduced biological activity in lentic ecosystems. Temperate lentic ecosystems experience a decrease in solar radiation during autumn and winter months that can limit the activity of phototrophic phytoplankton and bacteria (Bolsenga and Vanderploeg 1992). This reduced solar radiation may also reduce the rate of photochemical transformation of dissolved organic materials (Granéli et al. 1998), thus increasing the length of time over which these dissolved materials persist. Additionally, the cold autumn water temperatures may slow the rate of subsidy utilization through several mechanisms. First, cold temperatures reduce metabolic rates of most organisms and thus slow consumption and uptake rates among species (Loiterton et al. 2004). Many zooplankton initiate diapause in response to the reduced temperatures and light and food availability this time of year (Larsson and Wathne 2006), further reducing the ability of organisms to consume phytoplankton and bacteria. Finally, the high viscosity of cold water may decrease filtering rates of certain cladoceran zooplankton (Loiterton et al. 2004). The combined effect of these factors may slow the utilization of materials provided by autumn leaf subsidies until springtime conditions arrive. In addition to the effects of experiment location, it is important to recognize the differences between mesocosms and natural aquatic ecosystems (Dzialowski et al. 2014). For example, heat flux from pond sediments to surrounding cold waters can occur deep into the winter months (Kirillin et al. 2012), and has the potential to increase benthic biological activity over the limited overwinter activity in our mesocosms. Deeper ponds may also experience stronger winter physical stratification relative to our mesocosms (Bertilsson et al. 2013), and ponds with a different surface-to-volume ratio than our mesocosm may experience varied amount of total incoming solar radiation. Additionally, while our experiment focused on isolating long-term planktonic response to leaf subsidies, it would also be of

interest to examine how both benthic and pelagic consumers respond to leaf additions. Given the sensitivity of mesocosms to initial conditions and their tendency to experience increased variance through time (Dzialowski et al. 2014), our ability to detect treatment differences the spring following autumn leaf additions suggests a sustained effect of leaves on planktonic food webs. More generally, our research suggests that subsidies could be an underappreciated driver of pond phenology. Many ecological subsidies, including salmon spawning events (Schindler et al. 2013) or the fall of deciduous leaves, are important phenological events, in and of themselves. However, the ability of subsidies to also be drivers of phenology in other recipient ecosystems is less appreciated, despite evidence that subsidies across many systems may alter phenology. For example, freshly fallen leaf litter alters the development rates of the stonefly Tallaperla maria (Swan and Palmer 2006) and the emergence of aquatic insects affects the development of riparian spiders (Marczak and Richardson 2008). Similarly, our results show that subsidies can affect the subsequent timing of when physical, chemical, or biological response variables reach their maximum and minimum values during the year (e.g., peak chlorophyll was achieved 3 weeks earlier in control mesocosms during our experiment). In order to better explain inter-seasonal variation in phenological events across ecosystems, we suggest integrating the effects of ecological subsidies into the study of phenological events. In conclusion, it is unlikely that most subsidies are discrete resources, rather, certain ecological subsidies have the potential to link adjacent ecosystems long after the physical materials have been transferred. Our results highlight the importance of measuring beyond the immediate effects of how resource subsidies are integrated into their recipient food webs. Given the ability of anthropogenic disturbances to affect subsidies (Greig et al. 2012), our results suggest that deforestation and shifts in terrestrial community composition may alter aquatic ecosystems in unpredictable ways. Especially in light of the variation in which subsidies are seemingly utilized by surrounding biological communities in nature, we encourage exploring the long-term effects of subsidies across ecosystems to ensure that the contribution of subsidies to recipient food webs are accurately quantified. Author contribution statement  S. B. F. conceived the study; S. B. F., A. N. M., and K. L. C. designed research; S. B. F. and A. N. M. performed the experiment and analyzed the data; S. B. F. and A. N. M. wrote the first draft of the paper and K. L. C. contributed substantially to the revisions.

13

Oecologia Acknowledgments  We thank A. L. Ritger, J. V. Trout-Haney, and R. L. Wood for field and laboratory assistance and S. Stokoe for management of the Dartmouth Organic Farm. Comments from J. J. Gilbert, R. O. Hall, E. T. Irwin, and two anonymous reviewers greatly improved this manuscript. An Environmental Protection Agency STAR Fellowship and James S. McDonnell Complexity Postdoctoral Fellowship to S. B. F. and National Science Foundation grants DEB-1110369 to S. B. F. and K. L. C. and EF-0842267 to K. L. C., EF-0842112 to H. A. Ewing, and EF-0842125 to K. C. Weathers funded this research.

References Anderson N, Sedell J (1979) Detritus processing by macroinvertebrates in stream ecosystems. Annu Rev Entomol 24:351–377. doi:10.1146/annurev.en.24.010179.002031 Batzer DP, Palik BJ (2007) Variable response by aquatic invertebrates to experimental manipulations of leaf litter input into seasonal woodland ponds. Fundam Appl Limnol 168:155–162. doi:10.1127/1863-9135/2007/0168-0155 Baxter CV, Fausch KD, Carl Saunders W (2005) Tangled webs: reciprocal flows of invertebrate prey link streams and riparian zones. Freshwater Biol 50:201–220. doi:10.1111/j.1365-2427.2004. 01328.x Benfield EF, Webster JR, Tank JL, Hutchens JJ (2001) Long-term patterns in leaf breakdown in streams in response to watershed logging. Int Rev Hydrobiol 86:467–474. doi:10.1002/15222632(200107)86:4/5<467:AID-IROH467>3.0.CO;2-1 Berggren M, Ziegler SE, St-Gelais NF, Beisner BE, del Giorgio PA (2014) Contrasting patterns of allochthony among three major groups of crustacean zooplankton in boreal and temperate lakes. Ecology 95:1947–1959. doi:10.1890/13-0615.1 Bertilsson S, Burgin A, Carey CC, Fey SB, Grossart H, Grubisic LM, Jones ID, Kirillin G, Lennon JT, Shade A, Smyth RL (2013) The under-ice microbiome of seasonally frozen lakes. Limnol Oceanogr 58:1998–2012. doi:10.4319/lo.2013.58.6.1998 Bolsenga S, Vanderploeg H (1992) Estimating photosynthetically available radiation into open and ice-covered fresh-water lakes from surface characteristics—a high transmittance case study. Hydrobiologia 243:95–104. doi:10.1007/BF00007024 Brandstetter A, Sletten RS, Mentler A, Wenzel WW (1996) Estimating dissolved organic carbon in natural waters by UV absorbance (254 nm). Z Pflanzenernähr Bodenkd 159:605–607. doi:10.1002/ jpln.1996.3581590612 Cederholm C, Kunze M, Murota T, Sibatani A (1999) Pacific salmon carcasses: essential contributions of nutrients and energy for aquatic and terrestrial ecosystems. Fisheries 24:6–15. doi:10.1577/1548-8446(1999)024<0006:PSC>2.0.CO;2 Cole J, Carpenter S, Pace M, Van de Bogert M, Kitchell J, Hodgson J (2006) Differential support of lake food webs by three types of terrestrial organic carbon. Ecol Lett 9:558–568. doi:10.1111/j.1461-0248.2006.00898.x Cotrufo M, Ineson P (1996) Elevated CO2 reduces field decomposition rates of Betula pendula (Roth) leaf litter. Oecologia 106:525–530. doi:10.1007/BF00329711 Cottingham KL, Narayan L (2013) Subsidy quantity and recipient community structure mediate plankton responses to autumn leaf drop. Ecosphere 4:7 art89 Crumpton WG, Isenhart TM, Mitchell PD (1992) Nitrate and organic N analyses with second-derivative spectroscopy. Limnol Oceanogr 37:907–913. doi:10.4319/lo.1992.37.4.0907 Dzialowski AR, Rzepecki M, Kostrzewska-Szlakowska I, Kalinowska K, Palash A, Lennon JT (2014). Are the abiotic and biotic characteristics of aquatic mesocosms representative of in situ conditions? J Limnol 73 doi:10.4081/jlimnol.2014.721

13

Earl JE, Castello PO, Cohagen KE, Semlitsch RD (2014) Effects of subsidy quality on reciprocal subsidies: how leaf litter species changes frog biomass export. Oecologia 175:209–218. doi:10.1007/s00442-013-2870-x Elser JJ, Bracken MES, Cleland EE, Gruner DS, Harpole WS, Hillebrand H, Ngai JT, Seabloom EW, Shurin JB, Smith JE (2007) Global analysis of nitrogen and phosphorus limitation of primary producers in freshwater, marine and terrestrial ecosystems. Ecol Lett 10:1135–1142. doi:10.1111/j.1461-0248.2007.01113.x Fey SB, Cottingham KL (2012) Thermal sensitivity predicts the establishment success of non-native species in a mesocosm warming experiment. Ecology 93:2313–2320. doi:10.1890/12-0609.1 Fey SB, Mertens AN, Beversdorf LJ, McMahon KD, Cottingham KL (2015). Recognizing cross-ecosystem responses to changing temperatures: soil warming impacts pelagic food webs. Oikos doi:10.1111/oik.01939 France R, Peters R (1995) Predictive model of the effects on lake metabolism of decreased airborne litterfall through riparian deforestation. Conserv Biol 9:1578–1586. doi:10.1046/j.1523-1739.1995.09061578.x Granéli W, Lindell M, De Faria BM, de Assis Esteves F (1998) Photoproduction of dissolved inorganic carbon in temperate and tropical lakes–dependence on wavelength band and dissolved organic carbon concentration. Biogeochemistry 43:175–195. doi:10.102 3/A:1006042629565 Gratton C, Donaldson J, Zanden MJV (2008) Ecosystem linkages between lakes and the surrounding terrestrial landscape in northeast Iceland. Ecosystems 11:764–774 Greig HS, Kratina P, Thompson PL, Palen WJ, Richardson JS, Shurin JB (2012) Warming, eutrophication, and predator loss amplify subsidies between aquatic and terrestrial ecosystems. Glob Change Biol 18:504–514. doi:10.1111/j.1365-2486.2011.02540.x Guyette RP, Cole WG (1999) Age characteristics of coarse woody debris (Pinus strobus) in a lake littoral zone. Can J Fish Aquat Sci 56:496–505 Hieber M, Gessner MO (2002) Contribution of stream detrivores, fungi, and bacteria to leaf breakdown based on biomass estimates. Ecology 83:1026–1038 Higgs ND, Gates AR, Jones DOB (2014) Fish food in the deep sea: revisiting the role of large food-falls. PLoS One 9:e96016. doi:10.1371/journal.pone.0096016 Hongve D (1999) Production of dissolved organic carbon in forested catchments. J Hydrol 224:91–99. doi:10.1016/ S0022-1694(99)00132-8 Karlsson J, Bystrom P, Ask J, Ask P, Persson L, Jansson M (2009) Light limitation of nutrient-poor lake ecosystems. Nature 460:506–509. doi:10.1038/nature08179 Kelly PT, Solomon CT, Weidel BC, Jones SE (2014) Terrestrial carbon is a resource, but not a subsidy, for lake zooplankton. Ecology 95:1236–1242. doi:10.1890/13-1586.1 Kirillin G, Leppäranta M, Terzhevik A, Granin N, Bernhardt J, Engelhardt C, Efremova T, Golosov S, Palshin N, Sherstyankin P (2012) Physics of seasonally ice-covered lakes: a review. Aquat Sci 74:659–682 Larsson P, Wathne I (2006) Swim or rest during the winter—what is best for an alpine daphnid? Arch Hydrobiol 167:265–280. doi:10.1127/0003-9136/2006/0167-0265 Leech DM, Williamson CE (2000) Is tolerance to UV radiation in zooplankton related to body size, taxon, or lake transparency? Ecol Appl 10:1530–1540. doi:10.1890/1051-0761(2000)010[1530:ITT URI]2.0.CO;2 Lennon JT, Hamilton SK, Muscarella ME, Grandy AS, Wickings K, Jones SE (2013) A source of terrestrial organic carbon to investigate the browning of aquatic ecosystems. PLoS One 8:e75771 Loiterton B, Sundbom M, Vrede T (2004) Separating physical and physiological effects of temperature on zooplankton feeding rate. Aquat Sci 66:123–129. doi:10.1007/s00027-003-0668-3

Oecologia Marcarelli AM, Baxter CV, Mineau MM, Hall RO Jr (2011) Quantity and quality: unifying food web and ecosystem perspectives on the role of resource subsidies in freshwaters. Ecology 92:1215– 1225. doi:10.1890/10-2240.1 Marczak LB, Richardson JS (2008) Growth and development rates in a riparian spider are altered by asynchrony between the timing and amount of a resource subsidy. Oecologia 156:249–258 Nakano S, Murakami M (2001) Reciprocal subsidies: dynamic interdependence between terrestrial and aquatic food webs. Proc Natl Acad Sci USA 98:166–170. doi:10.1073/pnas.98.1.166 Odum E, Finn J, Franz E (1979) Perturbation-theory and the subsidystress gradient. Bioscience 29:349–352. doi:10.2307/1307690 Oertli B (1993) Leaf-litter processing and energy-flow through macroinvertebrates in a woodland pond (Switzerland). Oecologia 96:466–477. doi:10.1007/BF00320503 Polis G, Anderson W, Holt R (1997) Toward an integration of landscape and food web ecology: the dynamics of spatially subsidized food webs. Annu Rev Ecol Syst 28:289–316. doi:10.1146/ annurev.ecolsys.28.1.289 Pope R, Gordon A, Kaushik N (1999) Leaf litter colonization by invertebrates in the littoral zone of a small oligotrophic lake. Hydrobiologia 392:99–112. doi:10.1023/A:1003537232319 Porter K, Feig Y (1980) The use of DAPI for identifying and counting aquatic microflora. Limnol Oceanogr 25:943–948 Quinn G, Keough MJ (2002) Experimental design and data analysis for biologists. Cambridge University Press, Cambridge Richardson J (1992) Food, microhabitat, or both? macroinvertebrate use of leaf accumulations in a montane stream. Freshwater Biol 27:169–176. doi:10.1111/j.1365-2427.1992.tb00531.x Rubbo M, Kiesecker J (2004) Leaf litter composition and community structure: translating regional species changes into local dynamics. Ecology 85:2519–2525. doi:10.1890/03-0653 Rubbo MJ, Cole JJ, Kiesecker JM (2006) Terrestrial subsidies of organic carbon support net ecosystem production in temporary forest ponds: evidence from an ecosystem experiment. Ecosystems 9:1170–1176. doi:10.1007/s10021-005-0009-6 Rubbo MJ, Belden LK, Kiesecker JM (2008) Differential responses of aquatic consumers to variations in leaf-litter inputs. Hydrobiologia 605:37–44. doi:10.1007/s10750-008-9298-z Sabo J, Power M (2002) Numerical response of lizards to aquatic insects and short-term consequences for terrestrial prey. Ecology 83:3023–3036. doi:10.2307/3071839 Schindler DE, Armstrong JB, Bentley KT, Jankowski K, Lisi PJ, Payne LX (2013) Riding the crimson tide: mobile terrestrial consumers track phenological variation in spawning of an anadromous fish. Biol Lett 9:20130048. doi:10.1098/rsbl.2013.0048

Smith CR, Baco AR (2003) Ecology of whale falls at the deep-sea floor. Oceanogr Mar Biol Annu Rev 41:311–354 Swan C, Palmer M (2004) Leaf diversity alters litter breakdown in a piedmont stream. J N Am Benthol Soc 23:15–28. doi:10.1899/08 873593(2004)023<0015:LDALBI>2.0.CO;2 Swan C, Palmer M (2006) Composition of species leaf litter alters stream detritivore growth, feeding activity and leaf breakdown. Oecologia 147:469–478. doi:10.1007/s00442-005-0297-8 Takimoto G, Iwata T, Murakami M (2002) Seasonal subsidy stabilizes food web dynamics: balance in a heterogeneous landscape. Ecol Res 17:433–439. doi:10.1046/j.1440-1703.2002.00502.x Takimoto G, Iwata T, Murakami M (2009) Timescale hierarchy determines the indirect effects of fluctuating subsidy inputs on in situ resources. Am Nat 173:200–211. doi:10.1086/595759 Twiss MR, McKay RML, Bourbonniere RA, Bullerjahn GS, Carrick HJ, Smith REH, Winter JG, D’souza NA, Furey PC, Lashaway AR, Saxton MA, Wilhelm SW (2012) Diatoms abound in icecovered Lake Erie: an investigation of offshore winter limnology in Lake Erie over the period 2007–2010. J Great Lakes Res 38:18–30. doi:10.1016/j.jglr.2011.12.008 Ueveges V, Tapolczai K, Krienitz L, Padisak J (2012) Photosynthetic characteristics and physiological plasticity of an Aphanizomenon flos-aquae (Cyanobacteria, Nostocaceae) winter bloom in a deep oligo-mesotrophic lake (Lake Stechlin, Germany). Hydrobiologia 698:263–272. doi:10.1007/s10750-012-1103-3 Vanderploeg HA, Ludsin SA, Cavaletto JF, Hoeoek TO, Pothoven SA, Brandt SB, Liebig JR, Lang GA (2009) Hypoxic zones as habitat for zooplankton in Lake Erie: refuges from predation or exclusion zones? J Exp Mar Biol Ecol 381:S108–S120. doi:10.1016/j. jembe.2009.07.015 Wallace J, Eggert S, Meyer J, Webster J (1999) Effects of resource limitation on a detrital-based ecosystem. Ecol Monogr 69:409– 442. doi:10.1890/0012-9615(1999)069[0409:EORLOA]2.0.CO;2 Webster J, Meyer J (1997) Stream organic matter budgets. J N Am Benthol Soc 16:3–4. doi:10.2307/1468223 Welschmeyer NA (1994) Fluorometric analysis of chlorophyll-a in the presence of chlorophyll-b and pheopigments. Limnol Oceanogr 39:1985–1992 Wilkinson GM, Carpenter SR, Cole JJ, Pace ML, Yang C (2013) Terrestrial support of pelagic consumers: patterns and variability revealed by a multilake study. Freshwat Biol 58:2037–2049. doi:10.1111/fwb.12189 Williamson CE, Reid JW (2009) Copepoda. In: Thorp JH, Covich AP (eds) Ecology and classification of North American freshwater invertebrates. Academic Press, San Diego

13