Introduction

External nutrient imports are an important source for those habitats where abiotic conditions constrain internal nutrient cycling, weathering and biological fixation. Particularly deserts and polar regions can greatly benefit from external nutrient sources (e.g., Bokhorst et al. 2019; De la Peña Lastra 2020). Animals can act as an important transfer route of nutrients between biomes (Polis and Hurd 1996; Anderson and Polis 1999; Kolb et al. 2012; Doughty et al. 2016). In particular, birds forage in marine environments and deposit large amounts of nitrogen and phosphorus in the terrestrial environment (González-Bergonzoni et al. 2017; Zmudczyńska-Skarbek and Balazy 2017; Otero et al. 2018; De la Peña Lastra 2020), which typically results in increased primary production with higher nutritional value, buildup of soil organic layer, and enhanced microbial activity (Wright et al. 2010; Ellis et al. 2011; Zwolicki et al. 2013; Ball et al. 2015; Rampelotto et al. 2015). This in turn, can affect invertebrate populations which benefit through increased growth and reproduction (Zawierucha et al. 2015). Although many studies report positive effects of bird-derived nutrients on invertebrate abundance (Polis and Hurd 1996; Sanchez-Pinero and Polis 2000; Zmudczyńska et al. 2012; Bokhorst and Convey 2016; Bokhorst et al. 2019), there are also studies showing no effects (Gardner-Gee and Beggs 2009; Craig et al. 2012; Korobushkin and Saifutdinov 2019), or negative effects (Wright et al. 2010; Saifutdinov and Korobushkin 2020). The reported variation in invertebrate abundance in the presence of birds between studies may, among other factors, result from the quantity, quality and how the fecal matter is distributed as well as local climatic conditions and the invertebrate species studied. However, to date no comparative studies exist that address these issues.

Fecal nutritional content is largely dependent on food consumed (Emerson and Roark 2009; Spennemann and Watson 2017). As plant growth depends on various nutrients (Vitousek et al. 1993; Aerts and Chapin 2000), variations in nitrogen, phosphorous and other micro-nutrients of fecal matter from different bird taxa (Grant et al. 2022) may differently affect plant growth responses and associated invertebrates. Zwolicki et al. (2013) showed that a bird colony dominated by piscivorous species (Uria lomvia and Rissa tridactyla) increased soil available N and P more than a colony dominated by planktivorous birds (Alle alle). These contrasting bird fertilization effects resulted in a different vegetation composition, which can impact associated soil communities (Bezemer et al. 2010). Birds with a piscivorous diet are therefore, more likely to enhance invertebrate populations than those with a planktivorous diet. As bird species differ in their diet (Zwolicki et al. 2016; Pen-Mouratov and Dayan 2019) their impact through nutrient additions into the terrestrial biome may be species-specific.

Nesting site selection influences how guano is distributed which affects whether the nutrient loading is beneficial for plant growth or becomes toxic (Sanchez-Pinero and Polis 2000; Ellis 2005; Zelenskaya and Khoreva 2006; Kolb and Hambäck 2015). Larger spatial distribution, and dilution, of guano is found when birds nest in trees or on cliffs, while guano accumulates near surface-nesting birds. A small number of bird species burrow, resulting in improved soil nutrient mixing and promotes plant growth (Bancroft et al. 2005; Gardner-Gee and Beggs 2009). Invertebrate community responses may therefore, be dependent on the nest site selection of birds.

Invertebrate communities consist of several trophic feeding groups (Siepel and De Ruiter-Dijkman 1993; Schneider et al. 2004), and may therefore, not respond consistently to changes in primary production and resource quality resulting from bird-derived nutrients. Detritivores and primary consumers may benefit more than higher trophic levels from increased primary production (Wimp et al. 2010). However, saprophagous and detritivore beetle abundance increased in response to the presence of birds while the abundance of herbivorous and predatory beetles declined on islands in southern Sweden (Kolb et al. 2012). No responses were found between tardigrade feeding groups in response to bird-derived nutrients on arctic Svalbard (Zawierucha et al. 2019). Whether these community responses are mainly driven by species (taxa) specific responses or local climate is unclear. Considering the typical slow nutrient cycling and conservative plant traits in polar ecosystems (Reich and Oleksyn 2004; LeBauer and Treseder 2008), stronger invertebrate community responses can be expected to bird-derived nutrients than in milder climates.

This study aims (i) to identify to what extent bird taxa differ in their impact on terrestrial invertebrate abundance, (ii) determine if invertebrate taxa respond consistently to bird nutrient subsidies, and (iii) whether the effect of bird-derived nutrient subsidies on invertebrate communities is stronger in polar ecosystems? To address these questions, we performed an analysis of the published literature describing invertebrate responses associated with bird-derived nutrients. We hypothesize that: (1) the invertebrate response is dependent on the birds’ diet. (2) birds nesting off the ground have a stronger effect on invertebrate abundance than surface-nesting birds as guano is distributed over a larger area; (3) invertebrate primary consumers will respond stronger to bird-derived nutrients than higher trophic levels; and (4) bird-derived nutrients are more likely to increase invertebrate abundance in polar ecosystems. By addressing these hypotheses we aim to contribute to clarifying global and local patterns of terrestrial invertebrate communities in response to nutrient subsidies.

Methods

We identified data for this study from peer-reviewed literature searches from online libraries such as, Web of Science, Google Scholar and VU library resources, using multiple searches including the following search strings: ‘bird’, ‘invertebrate’, ‘nutrient’, ‘guano’ and various derivative terms such as: ‘seabird’, ‘beetle’, ‘ornithogenic’, ‘allochthonous’, ‘cormorant ‘, ‘springtail’, ‘arthropod’. We also screened cited references in papers and reviews to identify additional studies. The resulting papers were selected on the basis of quantitative data on invertebrate abundance in areas with birds present (treatment) and absent (control). This resulted in 50 papers published between 1977–2020 with study areas all over the world (see Fig. S1). From each of the 50 studies we obtained the following data: bird taxa, invertebrate abundance (species/family abundance was summed if only order was mentioned) and latitude. Bird diet and nesting site, and invertebrate trophic guilds were obtained from other resources when missing from the original study (Table 1). Nesting-site were categorized into: burrows, cliff-tree (representative of birds nesting above the ground), surface, mixed and ‘unknown’ if nesting-site selection was unclear.

Table 1 Number of available studies for comparison of invertebrate abundance response to bird-derived nutrients across bird and invertebrate taxa and climate zone
Full size table

Due to the large diversity of invertebrate groups represented across all studies, and lack of detailed knowledge on species involved, and their trophic status, we considered only predators and other guilds. Predators included Amblypgi, Araneae, Neuroptera, and Pseudoscorpionida. Other guilds included: Acari, Amphipoda, Annelida, Blattodea, Ciliophora, Coleoptera, Collembola, Diptera, Ephemeroptera, Gastropoda, Hemiptera, Hymenoptera, Isopoda, Lepidoptera, Myriapoda, Nematoda, Opiliones, Polyneoptera, Psocoptera, Rotifera, Tardigrada, Thysanoptera and Thysanura. Although this simplification does not do justice to the trophic complexity present in terrestrial invertebrate communities (Schneider et al. 2004), it does allow for comparisons between trophic levels. Some of the other guilds also include predatory species (e.g., Acari, Nematoda), but where these predators are present, they are part of a much larger community of primary consumers and therefore, unlikely to disproportionally able to affect the groups response within a specific study. In total there were 65 data points.

Comparison of bird effects on invertebrate abundance across climate zones was based on a simplified categorization across latitude; from 0 up to 23.5 was considered ‘Tropical’, 23.5 up to 40 ‘Subtropical or Dry’; 40 up to 60 ‘Temperate’; and above 60 ‘Polar’. This simplified approach was chosen, as refinement across higher resolution classifications required more studies than available.

Data analysis

To standardize response variables across studies we quantified the effect size of bird-derived nutrients compared to control (no birds present) using the following formula:

$$\frac{O-E}{E}$$

where ‘O’ represents the observed abundance in bird-affected sampling sites and ‘E’ represents the expected abundance based on control sites. This standardized bird-effect was quantified for each study and used to calculate mean response values. If the mean effect size was below 0, birds had a negative effect on the investigated invertebrate while values above 0 indicate positive effects. In addition, we calculated mean effect size through Hedges’ g, following the methods described in Lakens (2013). This analysis was based on the mean Hedges’ g from each of 44 studies, as not all publications (n = 50) provided N and SD. Chi-square testing was done on the proportion of studies with a positive or negative effect-size (H0: no. studies positive effect size = no. studies negative effect size). Wilcoxon signed rank tests were used to test if standardized bird-effects differed from 0. Kruskal–Wallis tests were used to identify significant (P < 0.05) differences between groups (bird taxa, diets, nesting site, invertebrate taxa, trophic guilds and climate zones). To identify if there were any significant trends in abundance response variables across latitude, we calculated pearson correlation coefficients. Log transformations of the standardized bird-effects were used for statistical analysis. Statistical analyses were performed using R version 4.2.3 (R-Core-Team 2023). We did not account for sampling strategy between studies, and assume that the approach and timing of this was optimized for each invertebrate group by the scientists involved.

Results

Bird effects

Forty of the 50 studies (80%) show a positive invertebrate response to the presence of birds (chi-square = 9.7, P = 0.002). On average invertebrate abundance was increased by 1027% across the whole data set (Table 2, Fig. 1), with a mean Hedges’ g value of 1.0 (sd = 2.3, n = 44). There were no significant differences among bird taxa (KW-chi-square = 8.2, df = 8, P = 0.418), their diets (KW-chi-square = 3.4, df = 6, P = 0.752) or nesting sites (KW-chi-square = 2.4, df = 4, P = 0.658) in their impact on invertebrate abundance (Table 2, Figs. 2, 3). However, mean effect-size among Charadriiformes was significantly higher than zero and this was also found for plankton- and mixed-diet, and birds nesting on cliffs and trees and a mixture of nesting places (Table 2).

Table 2 Mean bird-effect size on invertebrate abundance
Full size table
Fig. 1
figure 1

Bird effect on nearby invertebrate population size. Each triangle represents a study site (n = 50 see Table 1). Black cross indicates the overall mean effect. Grey circle represents the mean bird-effect (2.7, P = 0.03) excluding the study by Kennedy 1999 with an effect-size of 364 (data point not shown in figure)

Full size image
Fig. 2
figure 2

Bird-diet effect on invertebrate populations. Data is only shown for invertebrate groups with more than 13 data points. Box plots indicate the median standardised effect (O-E)/E values (error bars indicate the 95% confidence interval) and black crosses represent means. Invertebrate groups per study are presented by grey open triangles and outliers by black open circles. Values below zero indicate a negative effect and positive values for a positive effect on invertebrate abundance. Asterisk indicates a significant effect by bird-diet (P < 0.05 Wilcoxon signed-rank test) on invertebrate population size

Full size image
Fig. 3
figure 3

Bird nest site impact on invertebrate populations. Data is only shown for invertebrate groups with more than 13 data points. Box plots indicate the median standardised effect (O-E)/E values (error bars indicate the 95% confidence interval) and black crosses represent means. Invertebrate groups per study are presented by grey open triangles and outliers by black open circles. Values below zero indicate a negative effect and positive values for a positive effect on invertebrate abundance. Asterisk indicates a significant bird-effect (P < 0.05 Wilcoxon signed-rank test) on invertebrate population size

Full size image

Invertebrate responses

Invertebrate taxa response to the presence of birds was highly variable and the number of studies for each invertebrate taxon varied from 1 to 19 (Fig. 2, 3, and S6, Tables 1–2). There was no consistent difference in abundance response to the presence of birds between invertebrate taxa (KW-chi-square = 33.8, df = 26, P = 0.141) (Fig. S6). 14 out of 16 studies showed a positive effect-size for Coleoptera (chi-square = 9.0, P = 0.003), and similar patterns were found for Collembola (14 out of 19, chi-square = 4.3, P = 0.039) and Diptera (12 out of 14, chi-square = 7.1, P = 0.008), but not for Nematoda (11 out of 17, chi-square = 1.5, P = 0.225) (Fig S6). Araneae response was near significantly negative to bird presence (10 of 14, chi-square = 2.6, P = 0.109).

Coleoptera showed consistent significant positive abundance response to birds with a mixed diet, while there was a positive trend for Collembola (Fig. 2). Bird colonies including a mixture of nesting sites tend to have a consistent positive effect on Coleoptera abundance, while there were positive trends for Collembola and Diptera (Fig. 3). Tree or cliff nesting birds had positive trends for Collembola and Diptera (Fig. 3). On average, predators showed a low response (mean effect size = 0.865, sd = 51.9) to the presence of birds while that of other guilds was significantly higher (mean effect size = 10.4, sd = 2.51, P = 0.03, Fig. S5).

Climate and latitude effects

Bird presence effect on invertebrate abundance did not differ between climate zones (KW-χ2 = 1.3 df = 3, P = 0.726). No consistent significant trend was observed on the overall bird effect size across latitude (Fig. 4a). However, there were significantly more studies with a positive response in polar (16 out of 22; chi-square = 4.5; P = 0.033) and temperate (11 out of 14; chi-square = 4.6; P = 0.033) regions, while this was absent in studies from dry and tropical regions. Collembola and Diptera show trends of larger response to bird presence at higher latitudes than others. Piscivores bird-effects on invertebrate abundance was larger at higher latitudes (Fig. 4b) while the effect of birds feeding on a mixed diet was larger at lower latitudes (Fig. 4c). There was insufficient data to compare other bird diets across latitude.

Fig. 4
figure 4

Bird effect on invertebrate abundance across latitude. a Effect of bird presence on the abundance of Coleoptera, Collembola, Diptera and Nematoda at different latitudes. Latitudinal patterns of bird effect-sizes for birds on a fish diet b and mixed diets c Each data point represents a study in panel a, while points in panels b and c represent invertebrate taxa. Studies from the Southern Hemisphere are included as positive latitudes.

Full size image

Some species-specific responses

Some of the large variation in Collembola response was found among polar studies (Fig. 4). For instance; the mean Collembola response to the presence of birds was 0.34 for the study by Zmudczyńska-Skarbek et al. (2015) and 25.6 for Zmudczyńska-Skarbek et al. (2012) while both studies were done in the high arctic archipelago of Svalbard. The relative low effect-size was due to the limited response by some of the most dominant species (Folsomia quadrioculata, Anurida polaris and Hypogastrura tullbergi) whereas more than half of the present species showed a larger response (Fig. 5a). In the 2012 study all the dominant species showed a strong response to bird presence, although there was no consistent species response across those studies. The effect-size of birds on Collembola species was negatively correlated to Collembola species abundance, in the study by Zmudczyńska-Skarbek et al. (2015) (Fig. 5b).

Fig. 5
figure 5

Bird effect on Svalbard Collembola species. a species-specific response to bird presence with population densities of non-affected sites shown on the right, data from Zmudczyńska-Skarbek et al. (2015). b Correlation between Collembola population size and effect of bird presence on Collembola abundance. Each data point represents a Collembola species. Data from Zmudczyńska-Skarbek et al. (2015) and see panel (a)

Full size image

Discussion

One of the major challenges in identifying the effect of bird-derived nutrients on terrestrial invertebrate communities is that there are many interacting variables involved, such as the overall climate and environmental conditions, bird-diet and behaviour, changes in the vegetation composition, the invertebrates being studied and sampling design. Despite the fact that none of these aspects were consistently accounted for across the 50 studies, a significant positive invertebrate abundance response emerged. The positive effect-size is in contrast to the global meta-analyses on invertebrate responses to experimental N and P additions (Nessel et al. 2021).

Important differences between experimental nutrient additions and bird fecal matter is the composition of micro-elements, which are often lacking in the former case, and may result in nutrient imbalances with consequences for growth and reproductive output (Mitra and Flynn 2005). In addition, colonies may be considered ‘stable’ in time compared to experimental N-P additions ranged from 2–35 years (Nessel et al. 2021), which may affect invertebrate community assembly. Considering that climate change, pollution and anthropogenic activities are decreasing bird population sizes and their geographical distribution (Parmesan and Yohe 2003; Clucas et al. 2014; Trathan et al. 2015), large impacts can be expected for the terrestrial invertebrate communities that are affected by these birds.

Bird diet

The quantity and composition of nutrient additions can greatly influence plant nutritional content, growth rate, vegetation composition and therethrough, the associated microbiome (Guo et al. 2018; Mazei et al. 2018; Almela et al. 2022), and soil fauna (Pen-Mouratov and Dayan 2019). As many bird species differ in the food they consume, their fecal matter differs in nutritional content (Sitters et al. 2017) and as such, may influence the plant and animal communities near their nesting grounds (Zwolicki et al. 2013). However, there was only limited support for this hypothesis among the 50 studies within our analysis. Birds feeding on plankton had a significant positive impact on total invertebrate abundance while other diets showed much larger variation in effect-size. The nutritional content of the bird-fecal matter is unknown for the vast majority of the studies making it harder to identify whether this played a role. Piscivorous diets tend to enhance soil N and P more than planktivorous diets, resulting in different plant communities with altered nutrient status (Zwolicki et al. 2013, 2016). Data on the plankton-feeding birds was all derived from studies on Svalbard which may have affected the terrestrial invertebrate community response to this specific diet, and more studies in other regions are need to identify consistent patterns. In addition, other diet-effects turned out to be invertebrate-taxa specific. As bird diets have not been consistently compared within climate regions and across the same invertebrate groups, we cannot conclude whether this aspect of bird-effect plays a role on invertebrate communities. However, given that nutrient imbalances play a strong role in community composition across many biological groups (Busch and Phelan 1999; Huberty and Denno 2006; Cleveland and Liptzin 2007; Sitters et al. 2017; Aanderud et al. 2018) it seems highly likely that bird diet may turn out to be important for community assembly of terrestrial invertebrates.

Bird nesting

Distribution of bird fecal matter can greatly affect the nutrient concentration in the soil, which when deposited next to nests, as often found among surface nesting birds such as penguins (Lindeboom 1984), can lead to toxic levels that are detrimental for plant growth (Sanchez-Pinero and Polis 2000; Leishman and Wild 2001; Smykla et al. 2007; Crittenden et al. 2015; Natusch et al. 2017) and potentially result in lower invertebrate populations. However, there was no evidence for differences in effect-size between surface nesting birds compared to those nesting at higher elevation (in trees or on cliffs), which would redistribute nutrients across a larger ground area. Bird colonies with a mixture of nesting strategies tended to support larger populations of Coleoptera, Collembola and Diptera.

Changes in the soil structure and compaction due to the building of surface nests and trampling by birds seems likely (Gillham 1956; Ellis 2005), but little hard evidence was available for the 50 studies included in this comparison. More diverse nesting strategies within a bird colony are likely to create greater heterogeneity in nutrient distribution, although this requires specific testing, which may benefit terrestrial invertebrate communities.

Invertebrate taxa responses

Among the tested invertebrate taxa, Coleoptera, Collembola and Diptera showed near consistent positive effects to the presence of birds across studies, while other taxa showed greater proportions of negative effect-sizes. Invertebrate responses to experimental nutrient additions are often highly variable, group and context dependent (Scheu and Schaefer 1998; Konestabo et al. 2005; Roos et al. 2020; Sun et al. 2020), which to a large extent may explain the lack of consistent response observed here. Predatory species showed smaller responses to bird presence than lower trophic levels as the latter is more likely to directly benefit from changes in resource quality and primary production (Wimp et al. 2010). Various studies included different soil/vegetation substrates between sites with and without the presence of birds, resulting in a mean negative effect-size (e.g., Enríquez et al. 2018), and while this change in plant community plays a role in the overall ecosystem as affected by birds, it obscures direct impacts on invertebrate population sizes. Plants may affect invertebrates through resource provision but can also modify the microclimate with taller plants, resulting from additional nutrients, creating shaded cooler habitats (Green et al. 1984) which slows down mobility, growth and reproductive output of invertebrates (Nijssen et al. 2017). Changes in plant height and density, due to nutrient subsidies, can affect hiding places, thereby affecting predator–prey interactions with consequences for their relative populations (Crowder and Cooper 1982; Halaj et al. 2000). The extent to which plants do this, depends on the plant type and vegetation composition that is capable of growing near bird colonies (Ellis 2005). Plant community composition, plant stature and soil depth was not available for all studies and we could therefore, not account for this factor between studies. Some studies indicated that invertebrate abundance declines were directly related to too high nutrient levels (e.g., Kolb et al. 2010; 2015; Saifutdinov and Korobushkin 2020), which could have been accounted for by sampling at greater distance from bird colonies. Large effect-size variability may also be found at the invertebrate species level as not all species, within a taxonomic group, may respond to the presence of birds (Fig. 5). Other factors than food quality may be limiting in these bird-affected areas or that the carrying capacity of the habitat has been reached for the taxonomic group in question.

Climatic and latitude influences

Nutrient limitation plays a major role in shaping polar communities (Zwolicki et al. 2016; Bokhorst et al. 2019, 2022; Du et al. 2020), but we only observed increasing response trends, with large variability among the 22 polar studies (Fig. 4). Potential factors for the large invertebrate variability response have been discussed above. However, some trends exist across latitude for birds with a fish diet and those feeding on mixed sources, indicating that animal diet influencing ecosystem properties is climate dependent. Lower latitude ecosystems tend to be more P limited while N-limitation plays a stronger role at higher latitudes (Du et al. 2020). Although, we lack data on the specific nutrient content of the different bird taxa included in our comparison, diet does play a strong role in the nutrient composition of fecal matter (Levey and Karasov 1989; Adhurya et al. 2020). Fish eating birds seem to have a stronger positive effect on arthropod communities at higher latitudes (Fig. 4b) which may indicate that the ecosystems is benefitting from the additional N, whereas at lower latitudes, the mixed-feeding birds supply more P to the ecosystems. However, this hypothesis requires experimental testing across biomes (Bishop et al. 2010; Ott et al. 2014; Ball et al. 2018).

Conclusion

We were unable to provide a clear answer to the general question on how bird taxa differ in their impact on terrestrial invertebrate communities due to the unequal research focus across different bird taxa and the environments they inhabit. Other aspects, which were not addressed by any of the studies at hand, but may be relevant for terrestrial invertebrate responses include: (a) changes in the micro-climate due to altered vegetation composition and plant growth, (b) the transport and accumulation of toxic substances (e.g., heavy metals and micro-plastics) within bird colonies (Nie et al. 2012; Santamans et al. 2017; Hamilton et al. 2021; Grant et al. 2022), (c) physical changes in the soil structure due to nest building and trampling (Gillham 1956; Ellis 2005), (d) attraction of large vertebrate herbivores by the plant fertilization effect (Natusch et al. 2017) that alter plant community composition and soil structure and, (e) remains of bird feed, such as shell deposits and bird carcasses (Grant et al. 2022). The existence of nutrient gradients with distance to bird colonies (Erskine et al. 1998; Crittenden et al. 2015; Bokhorst et al. 2019) provides great natural settings to test ecological questions regarding community assembly rules and requires a spatially explicit sampling design.

The data from the 50 studies indicate that terrestrial invertebrate communities benefit from nutrient subsidies by birds. However, a direct link between enhanced nutrient availability and increased invertebrate population sizes, or nutrient transfer, using stable isotopes (e.g., Bokhorst et al 2019), was not always available. The large variation in effect-size between studies highlights the great complexity surrounding nutrient subsidies to communities (Hines et al. 2006; Subalusky and Post 2019; Benkwitt et al. 2021). Resolving these complexities will require systematic comparisons across space and time, between and across bird taxa and invertebrate species. Given the large role invertebrates play in various ecosystem processes, any changes in their population sizes, through climate change or anthropogenic influence (Carlini et al. 2007; Žydelis et al. 2009; Clucas et al. 2014), will impact on primary production, biodiversity and various ecosystem services. Considering that bird nutrient-subsidies are of a magnitude relevant for global nutrient modelling (Otero et al. 2018), their indirect consequences for terrestrial ecosystems and invertebrate communities may reach well beyond colony limits.