Advertisement

Polar Biology

, Volume 41, Issue 8, pp 1531–1545 | Cite as

The detritus-based microbial-invertebrate food web contributes disproportionately to carbon and nitrogen cycling in the Arctic

  • Amanda M. Koltz
  • Ashley Asmus
  • Laura Gough
  • Yamina Pressler
  • John C. Moore
Original Paper

Abstract

The Arctic is the world’s largest reservoir of soil organic carbon and understanding biogeochemical cycling in this region is critical due to the potential feedbacks on climate. However, our knowledge of carbon (C) and nitrogen (N) cycling in the Arctic is incomplete, as studies have focused on plants, detritus, and microbes but largely ignored their consumers. Here we construct a comprehensive Arctic food web based on functional groups of microbes (e.g., bacteria and fungi), protozoa, and invertebrates (community hereafter referred to as the invertebrate food web) residing in the soil, on the soil surface and within the plant canopy from an area of moist acidic tundra in northern Alaska. We used an energetic food web modeling framework to estimate C flow through the food web and group-specific rates of C and N cycling. We found that 99.6% of C processed by the invertebrate food web is derived from detrital resources (aka ‘brown’ energy channel), while 0.06% comes from the consumption of live plants (aka ‘green’ energy channel). This pattern is primarily driven by fungi, fungivorous invertebrates, and their predators within the soil and surface-dwelling communities (aka the fungal energy channel). Similarly, >99% of direct invertebrate contributions to C and N cycling originate from soil- and surface-dwelling microbes and their immediate consumers. Our findings demonstrate that invertebrates from within the fungal energy channel are major drivers of C and N cycling and that changes to their structure and composition are likely to impact nutrient dynamics within tundra ecosystems.

Keywords

Food web structure Energetic food web model Nutrient cycling C mineralization N mineralization Invertebrate Arctic Tundra 

Introduction

The Arctic is a major reservoir of global organic carbon (C) and considered an extremely important region in terms of its potential feedbacks to climate change (Schuur et al. 2008; Crowther et al. 2016). Of central concern is that climate warming is accelerating soil microbial respiration more than it is facilitating increased plant uptake of carbon dioxide (CO2) in this region, thereby transforming the Arctic from a sink to a source of atmospheric C. Plants and microbes are not the only actors in this system, as there is a diverse assemblage of consumers within the “green” (living plant-based) and “brown” (detritus-based) food webs. Interactions within and between these two food webs—predation, herbivory, detritivory—can directly affect the uptake, storage and mineralization of C and N (e.g., Bardgett and Wardle 2010). Arthropods and other invertebrates, including protozoans, comprise a large portion of the animal biomass on the tundra, outweighing their vertebrate counterparts by an order of magnitude by some estimates (Legagneux et al. 2012; Moore and deRuiter 2012). This suggests that as a group they may play an important role in affecting processes such as decomposition, primary production, and nutrient cycling.

The importance of invertebrates in regulating energy and nutrient flow through Arctic communities was recognized as early as 1923 by Summerhayes and Elton in their report on Bear Island in the Norwegian Svalbard archipelago (Summerhayes and Elton 1923). Yet almost 100 years later, while there have been numerous studies on the natural history of Arctic vertebrates and to a lesser extent invertebrates, few have addressed their influence on ecosystem structure and functioning in an integrated manner (see Oksanen et al. 1981). Notable exceptions include studies on the effects of vertebrate and soil invertebrate herbivores on plant communities (e.g., Gauthier et al. 2004; Gough et al. 2012; Sjögersten et al. 2012; Mosbacher et al. 2016) and the role of soil invertebrates, whose activity has been linked to C storage and N cycling in the tundra (e.g., Moore et al. 2003; Moore and deRuiter 2012; Sistla et al. 2013). In contrast, surface- and canopy-dwelling arthropods have typically been studied in terms of their role as a food source for birds (Legagneux et al. 2012; Bolduc et al. 2013; Boelman et al. 2015)—and not in terms of their impact on lower trophic levels or any particular ecosystem processes per se (but see Lund et al. 2017). This may be due in part to the low abundances of aboveground invertebrate herbivores (Danks 1992; Gelfgren 2010), which suggests that invertebrate consumption of plant biomass and effects on nutrient cycling are low within this system (Haukioja 1981). However, the proportionally higher abundances of detritivores as compared to herbivores suggest that the former may have a greater influence on C and N cycling than the latter (Ryan 1977).

Several frameworks have emerged that integrate aboveground and belowground assemblages, green and brown food webs, and the roles of active predators to study their impacts on ecosystem processes (e.g., Moore et al. 2003; Schmitz 2008a; Bardgett and Wardle 2010). For example, Bardgett and Wardle (2010) review how the activities and interactions of belowground communities may influence the productivity, diversity, and composition of plant communities. Moore et al. (2003) and Schmitz (2008a) focused on how belowground and aboveground invertebrate predators can impact biogeochemical cycling and other aspects of ecosystem functioning. These integrative frameworks seem appropriate to study Arctic systems, as several characteristics of Arctic invertebrate communities suggest that there should be strong links between the aboveground and belowground realms and between the green and brown components of the food web that are mediated by mobile arthropods. For example, many Arctic species are broad generalists that feed upon several different resources within their own and other trophic levels (Roslin et al. 2013; Wirta et al. 2015b). In particular, surface-dwelling generalist predators (e.g., wolf spiders and beetles) that serve as important links between the green and brown food webs in other herbaceous plant communities (Scheu 2001; Wardle 2002; Birkhofer et al. 2008; Bardgett and Wardle 2010), are extremely abundant in the Arctic tundra compared to most other groups (Høye and Forchhammer 2008; Gelfgren 2010; Wyant et al. 2011; Rich et al. 2013). Applying the integrated approach advocated above toward the study of Arctic invertebrate communities—and in particular to addressing the links between the green and brown food webs—would help us in identifying the role of these communities in regulating C and N dynamics in the tundra.

A comprehensive characterization of the invertebrate community would enable us to estimate the importance of these organisms in influencing ecological processes, and in turn, strengthen predictions about the response of this food web to disturbances such as climate change. To this end, we integrated samples from the soil, surface, and canopy habitats to build a complete microbial-invertebrate food web for the Arctic tundra (hereafter referred to as the invertebrate food web). Using these data, we then took an energetic food web modeling approach to quantify C and N flow through the web and estimate contributions by different functional feeding groups to respiration, N mineralization, and recycling of organic C and N. Based on previously published work in tundra ecosystems (e.g., Ryan 1977), we hypothesized that (a) organisms within the brown food web (i.e., invertebrates derived from detritus) would process the majority of energy and have a larger direct impact on C and N cycling than organisms within the green food web (i.e., those derived from live plant biomass). Given the large abundance of surface-dwelling generalist predators on the Alaskan tundra (Wyant et al. 2011; Rich et al. 2013), we also hypothesized that (b) these predators would serve as a major hub linking the green and brown food webs.

Methods

Study area

This study was performed in an area of moist acidic tundra on the North Slope of Alaska (68.952°N, 150.208°W, elev. 414 m), 37 km NW of Toolik Field Station and the Arctic Long-Term Ecological Research (LTER) site (Fig. 1). Moist acidic tundra (Bliss and Matveyeva 1992) is the dominant ecosystem type in this part of the Arctic, comprising more than 50% of the land area (Jandt et al. 2012). Mean soil pH at our site is 4.4 ± 0.2 (Bret-Harte et al. 2013) and mean annual temperature is 10 °C. Primary and secondary production are limited by the extremely short growing seasons, with mean temperatures above freezing for only three months (June, July, August) of the year (Hobbie et al. 2003). The plant community is characterized by mosses, dwarf evergreen shrubs, low-stature deciduous shrubs, and graminoids (Shaver and Chapin 1991; Bret-Harte et al. 2013). A comprehensive plant harvest in 2011 showed that total live plant biomass at our site was 1500 g m−2 and that aboveground net primary productivity (ANPP) was 200 g m−2 year−1, with over half of annual production being from graminoids (see Bret-Harte et al. 2013).
Fig. 1

Study site location on the North Slope of Alaska (68.952°N, 150.208°W, elev. 414 m), approximately 37 km NW of Toolik Field Station and the Arctic Long-Term Ecological Research (LTER) site. The site is characterized as moist acidic tundra, which is the dominant ecosystem type in this area of the Arctic

Sampling and processing of surface, canopy, and soil invertebrate communities

Differences in sizes and life history traits of organisms within the three microhabitats (i.e., soil, soil surface, and plant canopy) necessitated using a variety of sampling methods to characterize the entirety of the invertebrate food web. We established three parallel 50-m transects, adjacent to those used by Bret-Harte et al. (2013), separated by 5-m each for sampling the soil, surface, and canopy communities. Our samples were collected in July during peak plant biomass (July 18–25, 2013).

The canopy and surface arthropod communities were sampled every five meters along two of the three transects (total of ten samples each of canopy and surface). We sampled surface-dwelling organisms using pitfall traps comprised of clear cups (9 cm in diameter, 15 cm deep), filled ¼ full with a 75% ethanol solution. Traps were left out for one week (July 18-25), after which their contents were transferred to vials until further processing. Canopy-dwelling organisms were sampled on July 18, 2013 with a modified leaf vacuum in a 1-m2 area for 90 s at each location. Collected arthropods were placed in muslin bags and stored at −20 °C until sorted and then stored in 75% ethanol until identification. Canopy and surface-dwelling arthropods were identified to the family level using published keys (Triplehorn and Johnson 2005; Marshall 2006), with the exception of Collembola and Acari, which were identified only to subclass and any captured larvae, which were identified to order. We estimate that at this level of taxonomic resolution, our sampling methods detected roughly 82.5% of aboveground arthropod taxa (Online Resource 1).

For soil-dwelling organisms, we took soil samples every ten meters along the first transect (total of five samples) on July 18, 2013. Half of each soil sample was divided into 5–10 g subsamples to estimate the densities of bacteria, fungi, Protozoa, rotifers, tardigrades, enchytraeids, nematodes, and insects (larvae and adults). The other half of the sample was kept intact to estimate densities of soil-dwelling microarthropods and soil bulk density (see Gough et al. 2012 and Sistla et al. 2013). Samples were taken without regard to the plants present at each location along the sampling transect in order to capture the variability in vascular plant composition and associated roots present at our study site.

Fungal and bacterial densities were estimated from 5-g subsamples of soil using epiflourescent microscopy techniques (Bloem 1995). Fungi samples were stained with a calcoflour fluorescent brightener (see Frey et al. 1999) and read at 334-365 nm wavelength. Bacteria samples were stained with 5-(4,6 dichlorotriazin-2-yl) aminoflouorescein (DTAF) and read at 490 nm wavelength. Active fungal biomass was estimated as 10% of total fungal biomass (see Ingham and Klein 1984).

Protozoan (i.e., ciliates, flagellates, and amoeba) densities were estimated via the most probable number technique (Darbyshire et al. 1974), using a 10-g subsample of soil serially diluted with tenfold dilutions to 10−6 ml, and incubated at 14 °C with E. coli as a food source for 5 days. Population densities were estimated from distribution of presence and absence data across dilutions using the Most Probable Number Estimate Program (EPA 2013), the most common current approach for estimating soil protozoa biomass (Coleman et al. 2004; Crotty et al. 2012). Nematodes were extracted from 5 g of soil using the Baermann wet funnel technique (Baermann 1917). Isolated specimens were preserved using 10% formalin solution, counted and sorted into functional groups using compound microscopy. Enchytraeids, rotifers, tardigrades, and insect larvae were counted using a dissecting microscope from 5-g subsamples of soil immersed in deionized water.

The densities of soil-dwelling arthropods were estimated from the remaining half of the soil sample by 5-day heat-extraction into a solution of 90% ethanol and 10% glycerin using Tullgren funnels (Moore et al. 2000).

Microbial and invertebrate biomass estimates

Biomass estimates were obtained by multiplying the field estimates of population densities by taxon-specific estimates of the biomass of individuals. For surface and canopy insects, these estimates were based on allometric equations (Sample et al. 1993; Hódar 1997; Sabo et al. 2002; Gruner 2003) parameterized to the average body lengths (to 0.01 mm) of the first five individuals of a group encountered in each sample and the size of the sample area (see Pérez et al. 2016). The area sampled was explicit for canopy insects, set at 1 m2. For surface-dwelling arthropods caught by pitfall traps, we used the equation λ = Nt/(2×R×L) of Stoyan and Kushka (2001), where Nt is the average number of animals caught trap-day−1, R is the radius of the trap (0.0254 m), and L is the distance (m) a given animal group can walk in a day (personal observations and published movement estimates). Biomass values (mg C m−2) for soil-dwelling microbial and invertebrate functional groups (i.e., bacteria, fungi, collembola, enchytraeids, mites, nematodes, protozoa, rotifers, and tardigrades) were estimated from our density estimates and published information on mean individual dry weights (Hunt et al. 1987), corrected for soil bulk density.

Functional group assignment and food web structure

The food web was based on the biomass estimates and trophic interactions among functional groups of organisms and basal resources. Functional groups were based on primary food sources, feeding mode, habitat, and life history traits (Moore et al. 1988), which were determined from field and laboratory observations and published accounts (Online Resources 2, 3). All groups with the exception of biting flies were assumed to receive their energy from terrestrial sources. Based upon the known natural history of biting flies (and to a lesser extent non-biting midges), we assumed that this group relies substantially on aquatic resources acquired during the larval stage (i.e., diatoms and aquatic detritus) and to a lesser degree on blood meals (for reproduction) and nectar (to sustain flight activity) as adults (Danks 1992; Lundgren and Olesen 2005). Connectance (sensu May 1972) was estimated from the number of functional groups and basal resources (S) and trophic links (L) as C = 2L/(S×(S–1)). Additionally, in order to compare the trophic structure of each sub-web, we grouped functional groups into broader trophic groups (e.g., herbivore, predator, etc.) following conventions that have been used in previous tundra arthropod studies (e.g., Gelfgren 2010 and references therein; see Online Resource 2).

Energy fluxes and nutrient cycling

We simulated C fluxes between functional groups and rates of organic and inorganic C and N cycling for all functional groups using methods previously described for soil food webs (Hunt et al. 1987; de Ruiter et al. 1994; Moore and deRuiter 2012). Based upon the biomass estimates, this approach accounts for known death rates, feeding preferences, assimilation efficiencies, production efficiencies, and C:N ratios when deriving C and N cycling rates (mg C m−2 year−1; mg N m−2 year−1) for each functional group and the entire food web (see Hunt et al. 1987; de Ruiter et al. 1994). To do this, the model assumes that the system is at a steady state and that biomass production is equal to biomass loss from predation and natural death over a given unit of time. For any given trophic interaction, feeding rates, egestion rates, and mineralization rates of C and N are calculated as described by Moore and deRuiter (2012) and Andrés et al. (2016). We assumed that 50% of the estimated dry weight biomass of each group was C (Hunt et al. 1987; Doles 2000). Based on death rates and the assimilation efficiencies of each group, we calculated the total amount of organic C and N that would be recycled back to the system from the unassimilated biomass of prey (egestion—leavings, orts and feces) and the corpses of organisms that died non-predatory deaths (Zou et al. 2016). Estimates of inorganic C and N mineralization are based on the production efficiencies, assimilated consumption, and the C:N ratios of each functional feeding group.

The model accounts for both consumers with a single prey source and consumers with multiple prey sources by allowing feeding rates to depend upon the biomass of available prey. We assumed that soil-dwelling organisms primarily feed within the soil portion of the food web (Moore et al. 1988) but that there are some cross-feeding relationships between the soil and surface sub-webs and the surface and canopy sub-webs (e.g., generalist predators, biting flies; see Online Resource 3). Feeding preferences for a particular prey item within the same habitat were assigned a value of 1. For those surface and canopy-dwelling organisms with cross-feeding relationships, consumption of potential prey in other habitats or of prey that move between habitats (i.e., flies) were assumed to happen less frequently and were thus assigned preference values of either 0.1 or 0.9. For basal resources, we assumed that detritus, diatoms, lichen, moss, live plant biomass (roots, aboveground vascular plant tissue, pollen), and blood were not limiting resources. We assigned theoretical values of 2,500,000 mg C m−2 to detritus, 300,000 mg C m−2 to diatoms, and 300 mg C m−2 to all others. Under the steady state assumption, this does not affect the overall flux estimates but does allow us to estimate the fluxes from basal resources to consumers.

We ran 1000 simulations of the model to get an estimate of the variability across model runs. Each simulation had the same connectance and feeding preference matrices as input but different biomass estimates for each functional group. Specifically, due to the large variances and high coefficients of variations associated with the measured field estimates of biomass and the skewed nature of the distributions of these biomass estimates (i.e., often to the right), biomass estimates for each particular simulation were obtained by randomly sampling the gamma distribution of the biomass of each functional feeding group. The shape and scale of these gamma distributions were defined by the means and standard deviations of the biomass estimates from our field samples. Sampling from the gamma distribution obviates the problem of unrealistic negative values for biomass that would result from using a normal distribution and the need to create an arbitrary or ad hoc solution if a normal distribution were used (Bolker 2008). Thus, our simulations accounted for uncertainty in the absolute biomass contributions of the various groups to the entire community.

Sensitivity analysis

We tested the robustness of our model results by comparing them to those of other model runs that were based upon incomplete food webs and an additional food web without any specified feeding preferences (i.e., any potential prey was assigned a 1, whereas non-prey items were assigned 0 s). For the incomplete food webs, we removed each functional feeding group from the network, one at a time, while holding the rest of the food web constant. We reran the model 1000 times for each of these modified food webs (each time manipulating the biomass values of the remaining groups, as described above) and collected information on the total recycled organic C and N from egestion (i.e., leavings, orts, and feces) and corpses (due to natural deaths), total inorganic mineralized C and N, and total C consumed by the entire community (in mg C or N m−2 year−1).

Additionally, we explored the importance of including any given functional feeding group for the overall stability based on the diagonal strength (aka s-min) of the Jacobian matrix of the food web developed by de Ruiter et al. (1995) and Neutel et al. (2002). To calculate s-min the diagonal elements are based on the mass-specific natural (i.e., non-predatory) death multiplied by a constant ‘s’. Stability was estimated by determining the value of ‘s’ needed to ensure that the real parts of all the eigenvalues of the matrix are negative (e.g., Moore and Hunt 1988; de Ruiter et al. 1995; Rooney et al. 2006; Moore and deRuiter 2012). An s-min value of one indicates that the diagonal strength ensuring stability of the food web is dependent solely on the specific death rates of the functional groups. Hence low s-min values (s-min ≤ 1) indicate more stable food webs relative to those with high s-min (s-min ≥ 1).

Results

Food web structure and biomass

We identified 33 functional feeding groups across the entire food web, including samples from the soil, surface, and canopy habitats (Online Resource 2). Food web connectance was calculated as 0.32.

We calculated the dry biomass of the entire invertebrate food web as 15,161 mg m−2 (Table 1). Of this, 99.7% was comprised of microbial biomass (14,244 mg fungi m−2; 878 mg bacteria m−2). By considering the structure of the rest of the food web (i.e., excluding fungi and bacteria), soil-dwelling organisms made up the majority of biomass (45.1%), while surface and canopy communities comprised 23.8 and 31.1% of the biomass, respectively (Fig. 2). Biomass of soil-dwelling organisms was dominated by fungivores, which accounted for 75.6% of the biomass within the soil habitat (Table 1; Fig. 2); those fungivores with the most biomass included cryptostigmatid mites and collembola. Bactivores (amoebae, rotifers, nematodes), predators (ciliates, mites, nematodes, tardigrades), omnivores (nematodes), and other mixed-feeding microbivores contributed progressively decreasing amounts of biomass to the soil food web. Only 0.7% of the soil food web comprised belowground herbivores (phytophagous nematodes). Biomass within the community of surface-dwelling organisms was largely dominated by predators (95.7% of surface community), the majority of which was from predaceous beetles and wolf spiders. The canopy was the only habitat in which herbivores, primarily true bugs from families Cicadellidae and Delphacidae, comprised more than 1% of biomass within the local community (4.9% of canopy biomass). Detritivores, predators and parasitoids also each contributed 1.5, 0.7 and 0.8%, respectively, to canopy biomass (Table 1). More than 92% of the canopy biomass was comprised of hematophagous biting flies, 99.6% of which was from mosquitos. However, we acknowledge that this estimate of mosquito biomass is possibly inflated due to the collection method (i.e., attraction by mosquitoes to CO2 emitted by researchers during vacuum sampling).
Table 1

Functional feeding groups, biomass, and estimated rates of organic and inorganic C and N cycling by organisms sampled from the soil, surface, and canopy habitats in an area of moist acidic tundra in N. Alaska

Functional feeding group

Trophic group

Abbreviation in Fig. 2

Biomass

Non-predatory natural deaths

Unassimilated biomass of prey

Mineralized nutrients

(mg C m−2)

(mg C m−2 year−1)

(mg N m−2 year−1)

(mg C m−2 year−1)

(mg N m−2 year−1)

(mg CO2–C m−2 year−1)

(mg N m−2 year−1)

Amoebae

Bactivore

Amoebae

1.004 (1.076)

5.798 (0.195)

0.828 (0.027)

0.794 (0.026)

0.198 (0.006)

9.061 (0.306)

2.912 (0.098)

Bacteria

Bacteria

Bacteria

878.340 (698.041)

1003.807 (25.510)

250.951 (6.377)

0.000 (0.000)

0.000 (0.000)

2640.428 (59.813)

94.301 (2.136)

Beetles (predaceous)

Predator

PredBeetles

0.734 (2.582)

0.228 (0.028)

0.041 (0.005)

0.594 (0.072)

0.096 (0.012)

0.356 (0.043)

0.052 (0.006)

Cilliates

Predator

Cilliates

0.239 (0.169)

1.418 (0.030)

0.202 (0.004)

0.187 (0.004)

0.046 (0.001)

2.134 (0.046)

0.684 (0.014)

Collembola

 Soil-dwelling

Fungivore

SoilColl

6.604 (6.035)

26.469 (0.775)

3.308 (0.096)

95.445 (2.491)

9.544 (0.249)

62.039 (1.619)

5.368 (0.140)

 Surface-dwelling

Fungivore

SurfCollem

0.359 (0.251)

1.462 (0.031)

0.182 (0.003)

4.323 (0.094)

0.432 (0.009)

2.809 (0.061)

0.243 (0.005)

Enchytraeids

Microbivore

Enchy

1.735 (1.690)

8.397 (0.254)

1.679 (0.050)

63.743 (1.921)

6.377 (0.192)

12.748 (0.384)

0.426 (0.012)

Flagellates

Bactivore

Flagell

0.037 (0.032)

0.190 (0.005)

0.019 (0.000)

0.370 (0.012)

0.092 (0.003)

0.350 (0.011)

0.118 (0.003)

Flies

 Biting

Hematophage

BitingFlies

5.006 (0.833)

4.946 (0.026)

1.052 (0.005)

18.549 (0.100)

1.940 (0.010)

7.419 (0.040)

0.241 (0.001)

 Crane Flies

Detritivore

CraneFlies

0.001 (0.004)

0.000 (0.000)

0.000 (0.000)

0.000 (0.000)

0.000 (0.000)

0.000 (0.000)

0.000 (0.000)

 Herbivorous

Herbivore

HerbFlies

0.141 (0.106)

0.146 (0.003)

0.029 (0.000)

0.594 (0.015)

0.059 (0.001)

0.237 (0.006)

0.007 (0.000)

 Non-biting Midges

Detritivore

NonbitMidges

0.048 (0.078)

0.068 (0.003)

0.013 (0.000)

0.274 (0.013)

0.028 (0.001)

0.109 (0.005)

0.004 (0.000)

 Saprophagous

Detritivore

SapDiptera

0.225 (0.228)

0.343 (0.011)

0.070 (0.002)

1.367 (0.045)

0.136 (0.004)

0.546 (0.018)

0.016 (0.000)

Fungi

Fungi

Fungi

14244.236 (7889.573)

16907.239 (307.676)

1690.723 (30.767)

0.000 (0.000)

0.000 (0.000)

39457.050 (718.028)

3945.705 (71.802)

Lepidoptera

Herbivore

Lepid

0.070 (0.256)

0.014 (0.001)

0.002 (0.000)

0.072 (0.007)

0.007 (0.000)

0.029 (0.002)

0.002 (0.000)

Mites

 Cryptostigmatic

Fungivore

CrypMites

12.698 (16.296)

26.222 (1.031)

3.277 (0.128)

89.197 (3.228)

8.919 (0.322)

57.978 (2.098)

5.017 (0.181)

 Nematophagous

Predator

NemMites

0.252 (0.361)

0.466 (0.021)

0.058 (0.002)

1.920 (0.099)

0.192 (0.009)

1.248 (0.064)

0.108 (0.005)

 Non-cryptostigmatic

Fungivore

NonCrypMites

2.284 (1.885)

9.021 (0.234)

1.127 (0.029)

30.278 (0.790)

3.027 (0.079)

19.680 (0.513)

1.703 (0.044)

 Predatory

Predator

PredMites

1.038 (1.271)

1.866 (0.069)

0.233 (0.008)

4.918 (0.180)

0.594 (0.021)

4.795 (0.176)

0.568 (0.020)

Nematodes

 Bacteriophagous

Bactivore

BacNem

0.160 (0.154)

0.817 (0.025)

0.081 (0.002)

2.447 (0.107)

0.611 (0.026)

2.312 (0.101)

0.781 (0.034)

 Fungivorous

Fungivore

FungNem

0.049 (0.068)

0.187 (0.008)

0.018 (0.000)

1.438 (0.076)

0.143 (0.007)

0.555 (0.029)

0.055 (0.002)

 Omnivorous

Omnivore

OmniNem

1.735 (1.690)

13.988 (0.431)

1.398 (0.043)

31.992 (0.930)

7.987 (0.232)

30.232 (0.879)

10.205 (0.296)

 Phytophagous

Herbivore

PhytoNem

0.203 (0.308)

0.436 (0.020)

0.043 (0.002)

8.906 (0.673)

0.890 (0.067)

1.870 (0.141)

0.187 (0.014)

 Predatory

Predator

PredNem

0.127 (0.283)

0.674 (0.048)

0.067 (0.004)

2.611 (0.188)

0.649 (0.046)

1.645 (0.118)

0.553 (0.039)

Rotifers

Bactivore

Rotifers

0.253 (0.348)

1.566 (0.070)

0.156 (0.007)

2.832 (0.127)

0.708 (0.031)

2.677 (0.120)

0.904 (0.040)

Spiders

 Canopy, web-building

Predator

CanWebSpid

0.104 (0.122)

0.110 (0.003)

0.027 (0.000)

0.210 (0.007)

0.043 (0.001)

0.205 (0.007)

0.037 (0.001)

 Crab Spiders

Predator

CrabSpid

0.049 (0.174)

0.048 (0.006)

0.012 (0.001)

0.092 (0.011)

0.016 (0.002)

0.090 (0.010)

0.012 (0.001)

 Other surface-active

Predator

SurfSpid

0.350 (0.692)

0.075 (0.004)

0.018 (0.001)

0.158 (0.010)

0.027 (0.001)

0.154 (0.010)

0.021 (0.001)

 Surface, web-building

Predator

SurfWebSpid

0.093 (0.088)

0.094 (0.002)

0.023 (0.000)

0.203 (0.006)

0.028 (0.000)

0.198 (0.006)

0.016 (0.000)

 Wolf spiders

Predator

WolfSpid

2.163 (0.836)

1.080 (0.013)

0.270 (0.003)

2.304 (0.032)

0.393 (0.006)

2.247 (0.031)

0.287 (0.005)

Tardigrades

Predator

Tardig

0.203 (0.308)

1.530 (0.074)

0.153 (0.007)

3.668 (0.183)

0.913 (0.045)

3.466 (0.173)

1.166 (0.058)

True bugs

Herbivore

TrueBugs

0.690 (0.381)

0.676 (0.011)

0.107 (0.001)

2.813 (0.049)

0.281 (0.004)

1.125 (0.019)

0.068 (0.001)

Wasps (parasitic)

Parasitoid

Parasitoids

0.190 (0.219)

0.205 (0.007)

0.048 (0.001)

0.812 (0.029)

0.162 (0.005)

0.325 (0.011)

0.056 (0.002)

Total

15161.422

18019.600

1956.231

373.128

44.554

42326.13

4071.837

Total excluding microbes

38.846

108.554 s

14.556

373.128

44.554

228.655

31.831

Biomass is presented as the mean ± SE (in parentheses) mg C of the collected field samples. Surface and canopy data were combined by functional group for each set of paired plots (see sampling methods). Certain rare taxa that were typically associated with one habitat type that happened to be caught in another were disregarded. These included Acari, spiders from the family Thomisidae, and Coleoptera from the families Staphylinidae and Latridiidae that were caught in canopy samples, and Acari, Diptera, and spiders from the family Araneidae that were caught at the soil surface. Biomass estimates of Acari are from the soil sampling data only. Spiders from the family Linyphiidae use a variety of habitats, so biomass estimates were combined from the canopy and surface samples. Data on spider egg sacs were not included. Rates of C and N cycling for each functional feeding group are the mean ± SE of 1000 model simulations based on the biomass estimates. These rates include those from nutrients recycled back to the system from corpses of organisms that died non-predatory deaths and from the unassimilated biomass of prey (egestion—leavings, orts and feces), as well as rates of inorganic C and N mineralization. Biomass and rates of C and N cycling are expressed in mg C or N m−2 year−1

Fig. 2

Biomass estimates of the different trophic groups by habitat (canopy, soil surface, and soil) within the invertebrate food web in an area of moist acidic tundra of N. Alaska. The y-axis shows the total dry biomass in mg C m−2 of all organisms within the invertebrate food web by the contribution from each habitat. The x-axis indicates the proportion of biomass represented by each trophic group within the different habitats. Fungal and bacterial biomass estimates are not included here (see Table 1)

Food web energetics

Model results show that the majority of C flow within this food web is derived from detritus that enters the food web via consumption by fungi, bacteria, or detritivores (99.6% of total C flow; 60,686 mg C m−2 year−1), after which it is re-distributed among their respective consumers (Fig. 3; Online Resource 3). At the level of primary consumption, our estimates indicate that 15 times more C enters the detrital food web through the fungal channel (92.6% of total C flow; 56,360 mg C m−2 year−1) than through the bacterial channel (6.19%; 3772 mg C m−2 year−1). The large amount of C processed by the fungal energy channel in this tundra food web is also evidenced by the disproportionately larger biomass of fungivores when compared to bactivores in both the soil and surface habitats (biomass of fungivores 19× more than that of bactivores; Table 1; Fig. 2).
Fig. 3

Visualization of the energetic food web model of the invertebrate community in an area of moist acidic tundra in N. Alaska. Node sizes are proportional to the log-transformed average biomass (mg C m−2) of that functional feeding group (except for bacteria, fungi, and basal resources (i.e., roots, aboveground plant tissue, pollen, mammal blood, diatoms, detritus), whose node sizes were standardized due to their biomass estimates being too large to display comparatively with the other groups). Edges represent the feeding relationships between groups, and based on model results, edge widths are proportional to the amount of C transfer (mg C m−2 year−1) between these groups (except from detritus to bacteria and fungi, the quantities of which are much larger than shown here). See Table 1 for actual biomass estimates of the different functional feeding groups; model results of C flow rates between groups are contained in Online Resource 3. Nodes are color-coded by trophic group, which match those in Fig. 2. This figure was generated using the igraph package (Csardi and Nepusz 2006) in R with the LGL algorithm

Conversely, model results indicate that very little C enters the food web via herbivory (direct consumption of living primary production). Flow estimates of C from both belowground and aboveground herbivory made up only 0.06% of total energy flow (28 mg C m−2 year−1). As a consequence, estimated C flow from herbivores to predators is very small (3.04% of total C flow to predators; 1.30 mg C m−2 year−1). Rather, the majority of C flow to predators, particularly at the soil surface, originates from the detrital pool (85.05% of total C flow to predators; 36.48 mg C m−2 year−1). Cannibalism and intraguild predation are a substantial source of C for several of the predators as well (9.76% of total C flow to predators; 4.19 mg C m−2 year−1).

Contributions by invertebrate food web to C and N cycling

Estimates of C mineralization (i.e., CO2 respiration), N mineralization, unassimilated organic C and N, and organic C and N from the corpses of organisms that died natural deaths were also derived from the model for each functional feeding group (see Table 1 for means and standard errors across all model runs). Total C mineralization for the entire food web was estimated as 42,326 mg CO2–C m−2 year−1; total N mineralization was 4072 mg N m−2 year−1. Our estimates for the total organic C and N recycled back to the system by the natural death of individuals (non-predation events) were 18,020 mg C m−2 year−1 and 1956 mg N m−2 year−1.

The microbial community was responsible for the majority of this nutrient cycling (e.g., 99.4 and 99.2% of total C and N mineralization, respectively). Estimated rates of C and N cycling were all an order of magnitude higher in fungi than in bacteria (Table 1). Excluding microbes, total C mineralization, and N mineralization across the remainder of the food web were estimated as 229 mg C m−2 year−1 and 32 mg N m−2 year−1. Total organic C and N from egestion were 373 mg C m−2 year−1 and 44.5 mg N m−2 year−1, while recycled organic C and N from natural deaths were 109 and 15 mg m−2 year−1, respectively (Table 1). Soil-dwelling organisms were the largest contributors to all forms of C and N cycling. In particular, total respiration rates of soil-dwelling Collembola and cryptostigmatid mites were much higher than for any other animals. These groups were followed by the omnivorous nematodes and non-cryptostigmatid mites. In terms of N mineralization, omnivorous nematodes, soil-dwelling collembolan, cryptostigmatid mites, and amoebae were among the groups that contributed the most. In addition to non-cryptostigmatid mites, Enchytraeids, and biting flies, these same groups were also those that contributed the most organic C and N from egestion and from corpses due to non-predatory natural deaths (Table 1).

Sensitivity analysis

The results of our energetic food web model were robust to changes in the structure of the network and to changes in feeding preferences. Specifically, we found that excluding almost any functional feeding group from the network did not qualitatively change our main results that the majority of C entering the invertebrate food web is via the detrital pool and that among the invertebrates (i.e., not including microbes), soil-dwelling organisms are the largest contributors to C and N cycling (Online Resource 4). However, those models that excluded fungi or bacteria resulted in lower estimates of total C consumption, rates of total C and N mineralization, and rates of total C and N contributions from the corpses of non-predatory natural deaths (Online Resource 4, 5). Exclusion of fungi had the most drastic effect on estimates of C and N cycling, cutting the total C flow of the food web from 60,871 to 4676 mg C m−2 and drastically reducing the amount of organic and inorganic C and N contributed by the community (Online Resource 4, 5). Exclusion of several of the different soil-dwelling groups from the food web (e.g., soil-dwelling Collembola, cryptostigmatid and predatory mites, Enchytraeids) also resulted in lower estimates of total egested organic C and N, while excluding bacteria resulted in a higher estimate of egested organic C (Online Resource 4, 5).

The results from these additional models also indicated that the tundra invertebrate food web is highly stable. Exclusion of any functional feeding group or changes to the feeding preferences still resulted in all food webs having stable configurations (mean min-S values < 1; Online Resource 4). The food web showed particularly high stability when bacteria or soil-dwelling Collembola were excluded, while stability was lower when surface-dwelling spiders (e.g., wolf spiders and surface web spiders) were not included in the network (Online Resource 4).

Discussion

We characterized the structure of an Arctic invertebrate community by integrating data from the soil, soil surface, and canopy habitats and modeled C and N flow using an energetics-based food web model (Moore and deRuiter 2012). At our site in N. Alaska, we found that the majority of non-microbial biomass contained within the invertebrate food web is comprised of soil-dwelling organisms (45%), while surface and canopy communities contribute 24 and 31%, respectively. Consequently, invertebrate contributions to C and N cycling were also primarily derived from soil-dwelling organisms. Higher nutrient flow through the soil community was likely due in part to so much material entering the overall food web from detrital resources (99.6%) as opposed to from live plant biomass (0.06%). This pattern of uneven distribution of nutrient flow between the brown and green webs confirmed our first hypothesis and demonstrates that organisms reliant on detritus have a disproportionately larger impact on C and N cycling in tundra ecosystems than those reliant on live plant biomass.

Invertebrates within the brown food web process more nutrients than those within the green food web

Our model results confirmed that organisms within the brown food web—and not the green food web—process the majority of nutrients within this community. Furthermore, in accordance with findings from previous studies on soil food webs in tundra (e.g., Moore and William Hunt 1988; Moore et al. 2004; Rooney et al. 2006; Sistla et al. 2013), we found that detrital energy enters the brown food web primarily via fungi and their consumers (aka, the fungal energy channel), while the bacterial energy channel appears to play a smaller role in the breakdown of detrital matter. These results were robust to changes in the composition of the food web and to altering the feeding preferences within the network (Online Resources 4, 5). The large disparity in both biomass and C flow between the green and brown food webs suggests that overall, basal consumers within the brown food web, particularly the fungal community, are less constrained than herbivores in their ability to acquire energy from this system. Part of this difference may be due in part to some tundra herbivores having longer life spans and lower turnover rates than many of the soil-dwelling organisms (Strathdee and Bale 1998; Søvik et al. 2003). In addition, a portion of the brown food web is active year-round under the snow (Koltz unpublished, Clein and Schimel 1995; Zettel 2000), whereas arthropod consumers within the green food web, unlike their vertebrate counterparts, appear to only have access to plant resources during a very restricted window of the summer active season (Laperriere and Lent 1977; Huitu et al. 2003; Høye and Forchhammer 2008; Bolduc et al. 2013). A shorter active season among herbivores may be due to differences in overwintering strategies and cold hardiness between herbivores (e.g., MacLean 1983) and detritivores (Hodkinson et al. 1998). Regardless, such a short period of herbivore activity likely limits the amount of invertebrate biomass that can accumulate within the green food web.

The disproportionate contribution of the brown food web to C and N cycling suggests that changes to the structure of soil- or surface-dwelling communities will have much larger effects on ecosystem functioning than any changes within the canopy community. Additionally, while all simulated food webs were stable despite functional group exclusions (Online Resource 4), we observed greater changes in stability when soil-dwelling groups were excluded (especially bacteria and Collembola). These results suggest that soil organisms also play a more important role in maintaining food web stability relative to their aboveground counterparts. In particular, fungal-feeding detritivores (e.g., Collembola, cryptostigmatid and non-cryptostigmatid mites) process a large amount of C in this system (also see Rooney et al. 2006; Moore and deRuiter 2012; Sistla et al. 2013) and are known to be sensitive to changes in pH (van Straalen and Verhoef 1997), temperature (Coulson et al. 1996; Harte et al. 1996; Bokhorst et al. 2008; Day et al. 2009 ) and moisture (Verhoef and Selm 1983; Convey et al. 2003; Tsiafouli et al. 2005; Day et al. 2009). These groups and others have shown idiosyncratic responses to warmer temperatures and the associated lower soil moisture brought on by climate change (Hinzman et al. 2005) in the polar regions (e.g., Coulson et al. 1996; Koltz et al. unpublished; Nielsen and Wall 2013 and references therein), which could have consequences for C and N cycling and food web stability. Similarly, changes in plant community composition that affect the quality or quantity of litter inputs can influence the structure and composition of soil and surface-dwelling communities (Moore et al. 1988; Kaspari and Yanoviak 2009; Bardgett and Wardle 2010; Wyant et al. 2011). Shrub expansion, which is currently occurring in some areas of the Arctic (Myers-Smith et al. 2011), has been linked to changes in the composition of the surface-dwelling arthropod community (Rich et al. 2013) and a homogenization of soil food web structure (Sistla et al. 2013). The results of our analysis suggest that such community-level changes may have cascading effects; further replication of these methods across a variety of habitats would give us a better understanding of how variation in invertebrate food web structure may influence nutrient cycling and food web stability.

Consumption of live plant biomass by invertebrate herbivores is very small

Our model estimates that aboveground and belowground invertebrate herbivores on the tundra only consume 0.0019% of standing plant biomass annually. This is extremely low when compared to temperate and tropical systems where herbivorous arthropods can consume 1–35% of annual primary productivity (Schmidt and Kucera 1973; Coleman et al. 1976; Curry 1986; Detling 1988; Suzuki et al. 2013). Specifically, while total live plant biomass is approximately 1500 g m−2 at our study site (Bret-Harte et al. 2013), our model estimates show that invertebrate herbivores only consume roughly 0.028 g C m−2 of this biomass per year (Online Resource 3). These estimates are even lower than those from Devon Island in the High Arctic by Whitfield (1972), which indicated that invertebrate herbivores take ~1% of primary production. Overall our results suggest that consumptive effects of invertebrate herbivores are very small on the tundra, although model results by Wolf et al. (2008) and Barrio et al. (2017) suggest that these baseline levels may increase in the future with climate change. Rare herbivore outbreaks can also result in huge reductions to plant biomass (e.g., Pedersen and Post 2008; Lund et al. 2017). Such outbreaks are not known to occur in the Alaskan Arctic, but they have been documented in parts of Arctic Greenland (Pedersen and Post 2008; Lund et al. 2017) and are relatively common in the boreal forest (Volney and Fleming 2000; Soja et al. 2007) and in some areas of the Subarctic (Jepsen et al. 2008; Kaukonen et al. 2013). As treeline shifts northward and temperatures rise, these invertebrate herbivore outbreaks are expected to increase in intensity and frequency (Volney and Fleming 2000; Dale et al. 2001; Soja et al. 2007), with potential consequences for previously unaffected areas of the Arctic (Jepsen et al. 2011).

Surface-dwelling predators link the brown and green food webs

Our sampling showed that the majority of biomass at the soil surface consists of generalist predators (95.7% surface invertebrate biomass; Fig. 2). Consistent with our second hypothesis, model results indicated that these predators likely serve as an important link between the green and brown food webs (Fig. 3; Online Resource 3). The role of generalist predators in linking these food webs has been widely acknowledged in temperate ecosystems (e.g., Scheu 2001) but has received less attention in the Arctic. Our estimates indicate that among surface-dwelling predators, up to 27% of their energy resources may come from the canopy web while up to 46% may originate from the soil food web. Intraguild predation within and across habitats also appears to play an important role in sustaining predator populations (30% of C flow to surface predators). The generalist feeding behavior and high level of connectivity of surface-dwelling predators in this community may provide another potential explanation for the small amount of herbivore biomass. For example, soil-dwelling or intraguild prey may subsidize larger predator populations that are especially effective at keeping herbivore densities low (see Polis and Holt 1992; Schmitz 2008b). Our understanding of the role of these predators would benefit from further experimental and molecular work (e.g., Wirta et al. 2015a) that could confirm the strength of these feeding interactions, what proportion of prey comes from each sub-web, and how these interactions might vary seasonally and across habitats in the Arctic.

Whereas surface predators are the most important interface between green and brown energy pathways, flies with aquatic life cycles (midges, hematophagous flies, and others) serve as a bridge between aquatic and terrestrial systems in this food web (Dreyer et al. 2015). In particular, hematophagous flies (i.e., biting flies, especially mosquitoes) dominated the biomass of the canopy invertebrate assemblage and mobilized the largest fluxes of C and N of any group in the canopy food web (Fig. 3; Online Resource 3). Our sampling methodology likely resulted in an overestimate of biting fly density (abundance m−2), the majority of which was comprised of mosquitoes (99.6%). We stress that alongside other standardized methods of quantifying mosquito abundance (see Hoekman et al. 2016), an unbiased estimate of mosquito density would be a valuable tool for ecological accounting in the Arctic, particularly because we know that Arctic mosquitoes respond positively to warming (Culler et al. 2015). On the whole, however, our results are consistent with anecdotal observations of mosquito populations in Arctic regions (Danks 1992) and with previous studies documenting the primacy of flies in Arctic pollinator networks (Tiusanen et al. 2016).

Food web structure of the tundra invertebrate community

Our approach of integrating the communities of soil-, surface- and plant canopy-dwelling invertebrates enabled us to characterize this system from the perspective of the aboveground and belowground habitats, as well as the green and brown energy channels. Overall connectance for the invertebrate tundra food web was 32%, which is comparable to those of other published webs given the number of nodes (S = 33) within our food web (Briand 1983). Seasonal dynamics of different taxonomic groups vary throughout the Arctic summer, meaning that the structure of this invertebrate food web can also be quite variable (e.g., Høye and Forchhammer 2008). While our model simulations did incorporate aspects of community-level variability, more frequent sampling of the entire invertebrate community (e.g., soil, surface, and canopy-dwelling) throughout the active period would provide us with a better understanding of the seasonal variation in the structure and functional role of this food web.

Conclusions

Understanding the structure and function of the biological community is the first step in being able to predict how it might respond to disturbance. This study characterized the microbial-invertebrate food web in an Arctic tundra ecosystem, including soil, surface, and canopy-dwelling organisms. Our findings show that soil-dwelling organisms comprise the majority of biomass within this community and process more energy and nutrients than surface- or canopy-dwelling organisms. Our model results also indicate that invertebrate herbivores and their consumers that derive their energy from live plant matter (i.e., biota within the green food web) play a lesser role in processing nutrients and in mineralizing C and N on the tundra in comparison to those organisms derived from detrital resources. Consequently, changes to the structure and composition of the brown food web are likely to have a much greater impact on Arctic ecosystem functioning than any changes to the green food web. Given the sensitivity of many soil- and surface-dwelling organisms to changes in temperature and moisture, shifts in food web structure caused by climate change could have previously unforeseen consequences for C storage and nutrient cycling in the Arctic tundra. Future work will benefit from comparing the structure and energy flow of this food web and its potential role in nutrient cycling across the entire growing season, between different Arctic regions and habitats, and in response to climate change and other forms of disturbance.

Notes

Acknowledgements

We thank Gaius R. Shaver, Jim Laundre, and the Arctic LTER for support and coordinating transportation to the study area. We are also grateful to Greg Selby and Rod Simpson for assisting with the sampling and processing of soil samples and Sarah Meierotto, Kiki Contreras, Kathryn Daly, and PolarTREC teacher Nell Kemp for assistance processing the aboveground arthropod samples. Logistic support was provided by Toolik Field Station, University of Alaska, Fairbanks, USA and CH2MHILL; Fig. 1 was generated by the Toolik GIS Office. Funding for this research was provided by the U.S. National Science Foundation (OPP-0908602, 0909507, 0909441, and DEB 1026843 and 1210704), CREOi, and the National Geographic Committee for Research and Exploration.

Supplementary material

300_2017_2201_MOESM1_ESM.pdf (89 kb)
Supplementary material 1 (PDF 89 kb). Taxon rarefaction curve for surface and canopy communities sampled in July 2013 near Toolik Lake, Alaska. A total of 33 taxa were sampled; Estimates of extrapolated species richness suggest that the surface and canopy community actually contains 40 ± 7.1 taxa, indicating that we were able to capture roughly 82.5% of the aboveground arthropod community with our sampling methods and at this level of taxonomic resolution
300_2017_2201_MOESM2_ESM.pdf (42 kb)
Supplementary material 2 (PDF 41 kb) Designations of functional feeding and trophic groups for all arthropod families sampled from canopy and surface habitats. Trophic groups were used in reporting the biomass and trophic structure of each habitat type (see main text; Fig. 2) and functional feeding groups were used in the energetics-based food web model (Fig. 3; Online Resource 3)
300_2017_2201_MOESM3_ESM.xls (45 kb)
Supplementary material 3 (XLS 45 kb) Parameters used to initialize the energetics-based food web model and the simulated C flow rates between all consumer functional feeding groups within the invertebrate tundra food web. Included are estimates of the C:N ratio, death rate (DR), assimilation efficiency (AE), production efficiency (PE), and biomass (mean and standard deviation) for each functional feeding group. We assumed that detritus, diatoms, lichen, moss, live plant biomass (roots, vascular plants, pollen), and blood were not limiting resources and thus assigned theoretical values of 2,500,000 g C m−2 to detritus, 300,000 mg C m−2 to diatoms, and 300 mg C m−2 to all others. Estimates of C flow rates (mg C m−2 year−1) are from the complete (sampled) food web with assigned feeding preferences (see methods in main text). Zeroes denote no consumptive relationship between groups. Cross-habitat feeding relationships (e.g., between soil- and surface-dwelling organisms or surface- and canopy-dwelling organisms) are indicated by boldface type
300_2017_2201_MOESM4_ESM.xlsx (25 kb)
Supplementary material 4 (XLSX 25 kb). Summarized model results from the complete, sampled food web and all food web manipulations. Food web manipulations included not specifying feeding preferences and removing each sampled functional feeding group from the network, one at a time, while holding the rest of the food web constant. The results shown here are the mean and standard errors from 1000 model runs for each food web configuration. Estimates for total C flow and all rates of organic and inorganic C and N cycling are for the entire food web and expressed in mg C or N m−2 year−1. S-min is a measure of stability, estimated by determining the value of ‘s’ needed to ensure that the real parts of all the eigenvalues of the matrix are negative (e.g., Moore and Hunt 1988; de Ruiter et al. 1995; Rooney et al. 2006; Moore and deRuiter 2012). An s-min value of one indicates that the diagonal strength ensuring stability of the food web is dependent solely on the specific death rates of the functional groups. Hence low s-min values (s-min ≤ 1) indicate more stable food webs relative to those with high s-min (s-min ≥ 1)
300_2017_2201_MOESM5_ESM.pdf (1.4 mb)
Supplementary material 5 (PDF 1410 kb). Differences in the role of the invertebrate community in C consumption and cycling rates of organic and inorganic C and N between the complete, sampled food web versus those without feeding preferences or with individual functional feeding groups excluded (see “Methods” in main text)

References

  1. Andrés P et al (2016) Soil food web stability in response to grazing in a semi-arid prairie: the importance of soil textural heterogeneity. Soil Biol Biochem 97:131–143. doi: 10.1016/j.soilbio.2016.02.014 CrossRefGoogle Scholar
  2. Baermann G (1917) Eine eifache Methode Zur Auffindung con Anklyostomum (Nematoden) larvel in Erdproben Geneesk. Tijdschrift woor Nederlands Indie 57:131–137Google Scholar
  3. Bardgett RD, Wardle DA (2010) Aboveground-belowground linkages: biotic interactions, ecosystem processes, and global change. Oxford University Press, Oxford. doi: 10.1111/j.1442-9993.2012.02405.x CrossRefGoogle Scholar
  4. Barrio IC et al (2017) Background invertebrate herbivory on dwarf birch (Betula glandulosa-nana complex) increases with temperature and precipitation across the tundra biome. Polar Biol. doi: 10.1007/s00300-017-2139-7 CrossRefGoogle Scholar
  5. Birkhofer K, Wise DH, Scheu S (2008) Subsidy from the detrital food web, but not microhabitat complexity, affects the role of generalist predators in an aboveground herbivore food web. Oikos 117:494–500. doi: 10.1111/j.0030-1299.2008.16361.x CrossRefGoogle Scholar
  6. Bliss LC, Matveyeva NN (1992) Circumpolar arctic vegetation. In: Chapin FS III, Reynolds JF, Shaver GR, Svoboda J (eds) Arctic ecosystems in a changing climate: an ecophysiological perspective. Academic Press, San Diego, pp 59–89CrossRefGoogle Scholar
  7. Bloem J (1995) Fluorescent staining of microbes for total direct counts. In: Akkermans ADL, van Elsas JD, De Bruijn F (eds) Molecular microbial ecology manual. Springer, Netherlands, pp 367–378CrossRefGoogle Scholar
  8. Boelman NT et al (2015) Greater shrub dominance alters breeding habitat and food resources for migratory songbirds in Alaskan arctic tundra. Glob Change Biol 21:1508–1520. doi: 10.1111/gcb.12761 CrossRefGoogle Scholar
  9. Bokhorst S, Huiskes A, Convey P, Van Bodegom PM, Aerts R (2008) Climate change effects on soil arthropod communities from the Falkland Islands and the Maritime Antarctic. Soil Biol Biochem 40:1547–1556CrossRefGoogle Scholar
  10. Bolduc E et al (2013) Terrestrial arthropod abundance and phenology in the Canadian Arctic: modelling resource availability for Arctic-nesting insectivorous birds. Can Entomol 145:155–170. doi: 10.4039/tce.2013.4 CrossRefGoogle Scholar
  11. Bolker BM (2008) Ecological models and data in R. Princeton University Press, PrincetonGoogle Scholar
  12. Bret-Harte MS et al (2013) The response of Arctic vegetation and soils following an unusually severe tundra fire. Philos Trans Royal Soc B. doi: 10.1098/rstb.2012.0490 CrossRefGoogle Scholar
  13. Briand F (1983) Environmental Control of Food Web Structure. Ecol 64(2):253–263CrossRefGoogle Scholar
  14. Clein JS, Schimel JP (1995) Microbial activity of tundra and taiga soils at sub-zero temperatures. Soil Biol Biochem 27:1231–1234CrossRefGoogle Scholar
  15. Coleman D, Andrews R, Ellis J, Singh J (1976) Energy flow and partitioning in selected man-managed and natural ecosystems Agro-ecosystems 3:45–54Google Scholar
  16. Coleman DC, Crossley D, Hendrix PF (2004) Fundamentals of soil ecology. Academic press, CambridgeGoogle Scholar
  17. Convey P, Block W, Peat HJ (2003) Soil arthropods as indicators of water stress in Antarctic terrestrial habitats? Glob Change Biol 9:1718–1730CrossRefGoogle Scholar
  18. Coulson SJ et al (1996) Effects of experimental temperature elevation on high-arctic soil microarthropod populations. Polar Biol 16:147–153. doi: 10.1007/BF02390435 CrossRefGoogle Scholar
  19. Crotty FV, Adl SM, Blackshaw RP, Murray PJ (2012) Using stable isotopes to differentiate trophic feeding channels within soil food webs. J Eukaryot Microbiol 59:520–526PubMedCrossRefGoogle Scholar
  20. Crowther TW et al. (2016) Quantifying global soil carbon losses in response to warming. Nature 540:104–108 doi: 10.1038/nature20150. http://www.nature.com/nature/journal/v540/n7631/abs/nature20150.html—supplementary-information
  21. Csardi G, Nepusz T (2006) The igraph software package for complex network research. InterJournal Complex Syst 1695:1–9Google Scholar
  22. Culler LE, Ayres MP, Virginia RA (2015) In a warmer Arctic, mosquitoes avoid increased mortality from predators by growing faster. Proc Royal Soc B. doi: 10.1098/rspb.2015.1549 CrossRefGoogle Scholar
  23. Curry JP (1986) Above-ground arthropod fauna of four swedish cropping systems and its role in carbon and nitrogen cycling. J Appl Ecol 23:853–870. doi: 10.2307/2403939 CrossRefGoogle Scholar
  24. Dale VH et al (2001) Climate change and forest disturbances. BioScience 51:723–734. doi:10.1641/0006-3568(2001)051[0723:CCAFD]2.0.CO;2Google Scholar
  25. Danks HV (1992) Arctic Insects as Indicators of Environmental Change. Arctic 1992(45):159–166. doi: 10.14430/arctic1389 CrossRefGoogle Scholar
  26. Darbyshire J, Wheatley R, Greaves M, Inkson R (1974) A rapid micromethod for estimating bacterial and protozoan populations in soil. Revue d’Ecologie et de Biologie du Sol 11:465–475Google Scholar
  27. Day TA, Ruhland CT, Strauss SL, Park JH, Krieg ML, Krna MA, Bryant DM (2009) Response of plants and the dominant microarthropod, Cryptopygus antarcticus, to warming and contrasting precipitation regimes in Antarctic tundra. Global Change Biol 15:1640–1651CrossRefGoogle Scholar
  28. de Ruiter PC, Neutel A-M, Moore JC (1994) Modelling food webs and nutrient cycling in agro-ecosystems. Trends Ecol Evol 9:378–383. doi: 10.1016/0169-5347(94)90059-0 PubMedCrossRefGoogle Scholar
  29. de Ruiter PC, Neutel A-M, Moore JC (1995) Energetics, patterns of interaction strengths, and stability in real ecosystems. Science 269:1257PubMedCrossRefGoogle Scholar
  30. Detling JK (1988) Grasslands and savannas: regulation of energy flow and nutrient cycling by herbivores. In: Pomeroy LR, Alberts JJ (eds) Concepts of ecosystem ecology. Springer, Berlin, pp 131–148CrossRefGoogle Scholar
  31. Doles J (2000) A survey of soil biota in the arctic tundra and their role in mediating terrestrial nutrient cycling. University of Northern Colorado, GreeleyGoogle Scholar
  32. Dreyer J, Townsend PA, Iii JCH, Hoekman D, Vander Zanden MJ, Gratton C (2015) Quantifying aquatic insect deposition from lake to land. Ecology 96:499–509. doi: 10.1890/14-0704.1 PubMedCrossRefGoogle Scholar
  33. EPA US (2013) Most probably number (MPN) calculator version 2.0. In: User and system installation and administration manual. Environmental protection agency, Washington D.C., U.S., pp. 1–43Google Scholar
  34. Frey SD, Elliott ET, Paustian K (1999) Bacterial and fungal abundance and biomass in conventional and no-tillage agroecosystems along two climatic gradients. Soil Biol Biochem 31:573–585. doi: 10.1016/S0038-0717(98)00161-8 CrossRefGoogle Scholar
  35. Gauthier G, Bêty J, Giroux J-F, Rochefort L (2004) Trophic interactions in a high arctic snow goose colony. Integr Comp Biol 44:119–129. doi: 10.1093/icb/44.2.119 PubMedCrossRefGoogle Scholar
  36. Gelfgren M (2010) The importance of litter for interactions between terrestrial plants and invertebrates. Umea Universitet, UmeaGoogle Scholar
  37. Gough L, Moore JC, Shaver GR, Simpson RT, Johnson DR (2012) Above- and belowground responses of arctic tundra ecosystems to altered soil nutrients and mammalian herbivory. Ecology 93:1683–1694. doi: 10.1890/11-1631.1 PubMedCrossRefGoogle Scholar
  38. Gruner DS (2003) Regressions of length and width to predict arthropod biomass in the Hawaiian Islands. Pac Sci 57:325–336CrossRefGoogle Scholar
  39. Harte J, Rawa A, Price V (1996) Effects of manipulated soil microclimate on mesofaunal biomass and diversity. Soil Biol Biochem 28:313–322CrossRefGoogle Scholar
  40. Haukioja E (1981) Invertebrate herbivory at tundra sites Tundra ecosystems: a comparative analysis. Cambridge Univ Press, Cambridge, pp 547–555Google Scholar
  41. Hinzman L et al (2005) Evidence and implications of recent climate change in Northern Alaska and other arctic regions. Climatic Change 72:251–298. doi: 10.1007/s10584-005-5352-2 CrossRefGoogle Scholar
  42. Hobbie JE et al (2003) Climate forcing at the arctic LTER site. In: Greenland D (ed) Climate variability and ecosystem response at long-term ecological research (LTER) Sites. Oxford University Press, New York, pp 74–91Google Scholar
  43. Hódar JA (1997) The use of regression equations for estimation of prey length and biomass in diet studies of insectivore vertebrates. Miscel·lània Zoològica 20:1–10Google Scholar
  44. Hodkinson ID, Webb N, Bale J, Block W, Coulson S, Strathdee A (1998) Global change and Arctic ecosystems: conclusions and predictions from experiments with terrestrial invertebrates on Spitsbergen. Arctic Alpine Res 30:306–313CrossRefGoogle Scholar
  45. Hoekman D et al (2016) Design for mosquito abundance, diversity, and phenology sampling within the national ecological observatory network. Ecosphere 7:e01320. doi: 10.1002/ecs2.1320 CrossRefGoogle Scholar
  46. Høye T, Forchhammer M (2008) Phenology of high-arctic arthropods: effects of climate on spatial, seasonal and inter-annual variation Adv. Ecol Res 40:299–324Google Scholar
  47. Huitu O, Koivula M, Korpimäki E, Klemola T, Norrdahl K (2003) Winter food supply limits growth of northern vole populations in the absence of predation. Ecology 84:2108–2118CrossRefGoogle Scholar
  48. Hunt HW et al (1987) The detrital food web in a shortgrass prairie. Biol Fertil Soils 3:57–68. doi: 10.1007/bf00260580 CrossRefGoogle Scholar
  49. Ingham ER, Klein DA (1984) Soil fungi: relationships between hyphal activity and staining with fluorescein diacetate. Soil Biol Biochem 16:273–278. doi: 10.1016/0038-0717(84)90014-2 CrossRefGoogle Scholar
  50. Jandt RR, Miller EA, Yokel DA, Bret-Harte MS, Mack MC, Kolden CA (2012) Findings of Anaktuvuk River fire recovery study. US Bureau of Land Management, FairbanksGoogle Scholar
  51. Jepsen JU, Hagen SB, Ims RA, Yoccoz NG (2008) Climate change and outbreaks of the geometrids Operophtera brumata and Epirrita autumnata in subarctic birch forest: evidence of a recent outbreak range expansion. J Anim Ecol 77:257–264. doi: 10.1111/j.1365-2656.2007.01339.x PubMedCrossRefGoogle Scholar
  52. Jepsen JU, Kapari L, Hagen SB, Schott T, Vindstad OPL, Nilssen AC, Ims RA (2011) Rapid northwards expansion of a forest insect pest attributed to spring phenology matching with sub-Arctic birch. Glob Change Biol 17:2071–2083. doi: 10.1111/j.1365-2486.2010.02370.x CrossRefGoogle Scholar
  53. Kaspari M, Yanoviak SP (2009) Biogeochemistry and the structure of tropical brown food webs. Ecology 90:3342–3351. doi: 10.1890/08-1795.1 PubMedCrossRefGoogle Scholar
  54. Kaukonen M et al (2013) Moth herbivory enhances resource turnover in subarctic mountain birch forests? Ecology 94:267–272. doi: 10.1890/12-0917.1 PubMedCrossRefGoogle Scholar
  55. Laperriere AJ, Lent PC (1977) Caribou feeding sites in relation to snow characteristics in Northeastern Alaska. Arctic 30:101–108CrossRefGoogle Scholar
  56. Legagneux P et al (2012) Disentangling trophic relationships in a High Arctic tundra ecosystem through food web modeling. Ecology 93:1707–1716. doi: 10.1890/11-1973.1 PubMedCrossRefGoogle Scholar
  57. Lund M, Raundrup K, Westergaard-Nielsen A, López-Blanco E, Nymand J, Aastrup P (2017) Larval outbreaks in West Greenland: instant and subsequent effects on tundra ecosystem productivity and CO2 exchange. Ambio 46:26–38. doi: 10.1007/s13280-016-0863-9 PubMedPubMedCentralCrossRefGoogle Scholar
  58. Lundgren R, Olesen JM (2005) The Dense and highly connected world of Greenland’s plants and their pollinators Arctic. Antarctic Alpine Res 37:514–520. doi:10.1657/1523-0430(2005)037[0514:TDAHCW]2.0.CO;2Google Scholar
  59. MacLean SF Jr (1983) Life cycles and the distribution of psyllids (Homoptera) in arctic and subarctic Alaska. Oikos 40:445–451CrossRefGoogle Scholar
  60. Marshall SA (2006) Insects: their natural history and diversity: with a photographic guide to insects of eastern North America. Firefly Books Buffalo, New YorkGoogle Scholar
  61. May RM (1972) Will a large complex system be stable? Nature 238:413–414PubMedCrossRefGoogle Scholar
  62. Moore JC, deRuiter PC (2012) Energetic food webs: an analysis of real and model ecosystems. Oxford University Press, OxfordCrossRefGoogle Scholar
  63. Moore JC, William Hunt H (1988) Resource compartmentation and the stability of real ecosystems. Nature 333:261–263CrossRefGoogle Scholar
  64. Moore JC, Walter DE, Hunt HW (1988) Arthropod regulation of micro- and mesobiota in below-ground detrital food webs. Annu Rev Entomol 33:419–435. doi: 10.1146/annurev.en.33.010188.002223 CrossRefGoogle Scholar
  65. Moore JC, Tripp BB, Simpson RT, Coleman DC (2000) Springtails in the classroom: collembola as model organisms for inquiry-based laboratories. Am Biol Teacher 62:512–519Google Scholar
  66. Moore JC, McCann K, Setälä H, De Ruiter PC (2003) Top-down is bottom-up: does predation in the rhizosphere regulate aboveground dynamics? Ecology 84:846–857. doi:10.1890/0012-9658(2003)084[0846:TIBDPI]2.0.CO;2Google Scholar
  67. Moore JC et al (2004) Detritus, trophic dynamics and biodiversity. Ecol Lett 7:584–600. doi: 10.1111/j.1461-0248.2004.00606.x CrossRefGoogle Scholar
  68. Mosbacher JB, Kristensen DK, Michelsen A, Stelvig M, Schmidt NM (2016) Quantifying Muskox plant biomass removal and spatial relocation of nitrogen in a high Arctic Tundra ecosystem Arctic. Antarctic Alpine Res 48:229–240. doi: 10.1657/AAAR0015-034 CrossRefGoogle Scholar
  69. Myers-Smith IH et al (2011) Shrub expansion in tundra ecosystems: dynamics, impacts and research priorities. Environ Res Lett 6:045509CrossRefGoogle Scholar
  70. Neutel A-M, Heesterbeek JA, de Ruiter PC (2002) Stability in real food webs: weak links in long loops. Science 296:1120–1123PubMedCrossRefGoogle Scholar
  71. Nielsen UN, Wall DH (2013) The future of soil invertebrate communities in polar regions: different climate change responses in the Arctic and Antarctic? Ecol Lett 16:409–419. doi: 10.1111/ele.12058 PubMedCrossRefGoogle Scholar
  72. Oksanen L, Fretwell SD, Arruda J, Niemela P (1981) Exploitation ecosystems in gradients of primary productivity. Am Nat 118:240–261CrossRefGoogle Scholar
  73. Pedersen C, Post E (2008) Interactions between herbivory and warming in aboveground biomass production of arctic vegetation. BMC Ecol. doi: 10.1186/1472-6785-8-17 PubMedPubMedCentralCrossRefGoogle Scholar
  74. Pérez JH et al (2016) Nestling growth rates in relation to food abundance and weather in the Arctic. Auk 133:261–272. doi: 10.1642/AUK-15-111.1 CrossRefGoogle Scholar
  75. Polis GA, Holt RD (1992) Intraguild predation: the dynamics of complex trophic interactions. Trends Ecol Evol 7:151–154PubMedCrossRefGoogle Scholar
  76. Rich ME, Gough L, Boelman NT (2013) Arctic arthropod assemblages in habitats of differing shrub dominance. Ecography 36:994–1003. doi: 10.1111/j.1600-0587.2012.00078.x CrossRefGoogle Scholar
  77. Rooney N, McCann K, Gellner G, Moore JC (2006) Structural asymmetry and the stability of diverse food webs. Nature 442:265–269. http://www.nature.com/nature/journal/v442/n7100/suppinfo/nature04887_S1.html
  78. Roslin T, Wirta H, Hopkins T, Hardwick B, Várkonyi G (2013) Indirect Interactions in the High Arctic. PLoS ONE 8:e67367. doi: 10.1371/journal.pone.0067367 PubMedPubMedCentralCrossRefGoogle Scholar
  79. Ryan JK (1977) Synthesis of energy flows and population dynamics of Truelove Lowland invertebrates [Insects, protozoa, nematodes]. In: Bliss LC (ed) Truelove Lowland, Devon Island, Canada: a High Arctic Ecosystem. The University of Alberta Press, Edmonton, pp 325–346Google Scholar
  80. Sabo JL, Bastow JL, Power ME (2002) Length–mass relationships for adult aquatic and terrestrial invertebrates in a California watershed. J North Am Benthol Soc 21:336–343. doi: 10.2307/1468420 CrossRefGoogle Scholar
  81. Sample BE, Cooper RJ, Greer RD, Whitmore RC (1993) Estimation of insect biomass by length and width. Am Midland Nat 129:234–240. doi: 10.2307/2426503 CrossRefGoogle Scholar
  82. Scheu S (2001) Plants and generalist predators as links between the below-ground and above-ground system. Basic Appl Ecol 2:3–13. doi: 10.1078/1439-1791-00031 CrossRefGoogle Scholar
  83. Schmidt ND, Kucera C (1973) Arthropod food chain energetics in a Missouri tall grass prairie. University of Missouri, ColumbiaGoogle Scholar
  84. Schmitz OJ (2008a) Effects of predator hunting mode on grassland ecosystem function. Science 319:952–954. doi: 10.1126/science.1152355 PubMedCrossRefGoogle Scholar
  85. Schmitz OJ (2008b) Herbivory from individuals to ecosystems Annual Review of Ecology. Evol Syst 39:133–152CrossRefGoogle Scholar
  86. Schuur EAG et al (2008) Vulnerability of permafrost carbon to climate change: implications for the global carbon cycle. Bioscience 58:701–714. doi: 10.1641/b580807 CrossRefGoogle Scholar
  87. Shaver GR, Chapin FS (1991) Production: biomass relationships and element cycling in contrasting arctic vegetation types. Ecol Monogr 61:1–31. doi: 10.2307/1942997 CrossRefGoogle Scholar
  88. Sistla SA, Moore JC, Simpson RT, Gough L, Shaver GR, Schimel JP (2013) Long-term warming restructures Arctic tundra without changing net soil carbon storage. Nature 497:615–618 doi: 10.1038/nature12129. http://www.nature.com/nature/journal/v497/n7451/abs/nature12129.html - supplementary-information
  89. Sjögersten S, van der Wal R, Woodin S (2012) Impacts of grazing and climate warming on C pools and decomposition rates in arctic environments. Ecosystems 15:349–362. doi: 10.1007/s10021-011-9514-y CrossRefGoogle Scholar
  90. Soja AJ et al (2007) Climate-induced boreal forest change: predictions versus current observations. Global Planet Change 56:274–296. doi: 10.1016/j.gloplacha.2006.07.028 CrossRefGoogle Scholar
  91. Søvik G, Leinaas HP, Ims RA, Solhøy T (2003) Population dynamics and life history of the oribatid mite Ameronothrus lineatus (Acari, Oribatida) on the high arctic archipelago of Svalbard. Pedobiologia 47:257–271. doi: 10.1078/0031-4056-00189 CrossRefGoogle Scholar
  92. Stoyan D, Kushka V (2001) On animal abundance estimation based on pitfall traps. Biom J 43:45–52CrossRefGoogle Scholar
  93. Strathdee A, Bale J (1998) Life on the edge: insect ecology in arctic environments. Annu Rev Entomol 43:85–106PubMedCrossRefGoogle Scholar
  94. Summerhayes VS, Elton CS (1923) Bear Island. J Ecol 11:216–233. doi: 10.2307/2255864 CrossRefGoogle Scholar
  95. Suzuki S, Kitayama K, S-i Aiba, Takyu M, Kikuzawa K (2013) Annual leaf loss caused by folivorous insects in tropical rain forests on Mt. Kinabalu, Borneo. J For Res 18:353–360. doi: 10.1007/s10310-012-0356-z CrossRefGoogle Scholar
  96. Tiusanen M, Hebert PDN, Schmidt NM, Roslin T (2016) One fly to rule them all—muscid flies are the key pollinators in the Arctic. Proc Royal Soc B. doi: 10.1098/rspb.2016.1271 CrossRefGoogle Scholar
  97. Triplehorn CA, Johnson NF (2005) Borror and DeLong’s Introduction to the Study of Insects, 7th edn. Thomson Brooks/Cole, BelmontGoogle Scholar
  98. Tsiafouli MA, Kallimanis AS, Katana E, Stamou GP (2005) &, Sgardelis SP. Responses of soil microarthropods to experimental short-term manipulations of soil moisture Applied Soil Ecology 29:17–26Google Scholar
  99. van Straalen NM, Verhoef HA (1997) The development of a bioindicator system for soil acidity based on arthropod pH preferences. J Appl Ecol 34:217–232. doi: 10.2307/2404860 CrossRefGoogle Scholar
  100. Verhoef HA, Selm AJV (1983) Distribution and population dynamics of collembola in relation to soil moisture holarctic. Ecology 6:387–394. doi: 10.2307/3682436 CrossRefGoogle Scholar
  101. Volney WJA, Fleming RA (2000) Climate change and impacts of boreal forest insects agriculture. Ecosyst Environ 82:283–294. doi: 10.1016/S0167-8809(00)00232-2 CrossRefGoogle Scholar
  102. Wardle DA (2002) Communities and ecosystems: linking the aboveground and belowground components, vol 34. Princeton University Press, PrincetonGoogle Scholar
  103. Whitfield D (1972) Systems analysis Devon Island IBP project, high Arctic ecosystem Dept Botany. Univ Alberta, Edmonton, pp 392–409Google Scholar
  104. Wirta HK et al (2015a) Exposing the structure of an Arctic food web. Ecol Evol 5:3842–3856. doi: 10.1002/ece3.1647 PubMedPubMedCentralCrossRefGoogle Scholar
  105. Wirta HK, Weingartner E, Hambäck PA, Roslin T (2015b) Extensive niche overlap among the dominant arthropod predators of the High Arctic. Basic Appl Ecol 16:86–92. doi: 10.1016/j.baae.2014.11.003 CrossRefGoogle Scholar
  106. Wolf A, Kozlov M, Callaghan T (2008) Impact of non-outbreak insect damage on vegetation in northern Europe will be greater than expected during a changing climate. Climatic Change 87:91–106. doi: 10.1007/s10584-007-9340-6 CrossRefGoogle Scholar
  107. Wyant KA, Draney ML, Moore JC (2011) Epigeal Spider (Araneae) Communities in Moist Acidic and Dry Heath Tundra at Toolik Lake, Alaska. Arctic Antarctic Alpine Res 43:301–312. doi: 10.1657/1938-4246-43.2.301 CrossRefGoogle Scholar
  108. Zettel J (2000) Alpine Collembola - adaptations and strategies for survival in harsh environments Zool-Anal. Complex Syst 102:73–89Google Scholar
  109. Zou K, Thébault E, Lacroix G, Barot S (2016) Interactions between the green and brown food web determine ecosystem functioning. Funct Ecol 30:1454–1465. doi: 10.1111/1365-2435.12626 CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany 2017

Authors and Affiliations

  • Amanda M. Koltz
    • 1
  • Ashley Asmus
    • 2
  • Laura Gough
    • 3
  • Yamina Pressler
    • 4
  • John C. Moore
    • 4
    • 5
  1. 1.Department of BiologyWashington University in St. LouisSt. LouisUSA
  2. 2.Department of BiologyUniversity of Texas at ArlingtonArlingtonUSA
  3. 3.Department of Biological SciencesTowson UniversityTowsonUSA
  4. 4.Natural Resource Ecology LaboratoryColorado State UniversityFt. CollinsUSA
  5. 5.Department of Ecosystem Science and SustainabilityColorado State UniversityFt. CollinsUSA

Personalised recommendations