Foraging habits of spring migrating waterfowl in the upper mississippi river and great lakes region
Yüklə 0,59 Mb.
|
Waterfowl Collection To estimate diet during spring migration, I collected foraging female mallards, gadwall, blue-winged teal, scaup, and ring-necked ducks with a shotgun. Collection began as soon as ice thawed and continued until migrant ducks vacated the study areas (early May). I attempted to collect only individuals that had fed for ≥ 10 minutes to ensure birds contained ingesta. In some cases, dense vegetation reduced visibility (i.e., forested and emergent wetlands), and I only collected individuals that I knew had been in the habitat for an extended time and were suspected to have been feeding. I collected foraging females using a layout boat, by stalking, or from camouflaged observation blinds. Layout boats were operated with a trolling motor and camouflaged with sheets of artificial grass. I approached ducks in the layout boat from upwind to encourage them to flush in the direction of the collector. I recorded locations of collected birds with a global positioning system (GPS) unit and created a shapefile containing the collection data using a handheld personal digital assistant (PDA). Immediately following collection, I injected the esophagi with 10% buffered formalin solution to prevent post-mortem digestion of food items (Swanson and Bartonek 1970) and placed a zip-tie at the base of the skull to ensure formalin and food items were retained. I assigned ducks an identification number, placed an identification tag on their leg, and refrigerated them until the esophageal tract and proventriculus could be removed. I removed the esophageal tract and proventriculus within 5 days of collection and stored them in vials of 10% buffered formalin solution marked with the unique bird number and species. Laboratory Analysis Esophagi of collected ducks were analyzed at Southern Illinois University Carbondale’s (SIUC) Cooperative Wildlife Research Laboratory Annex. To determine diet, I removed, rinsed, and sorted contents of the esophagus and proventriculus and used a dissecting microscope to separate animal and plant food items. Animal food item identification was conducted at SIUC and seed identification was conducted at The Ohio State University. Animal foods recovered from esophageal contents were identified to family (Merritt and Cummins 1996), whereas plant material was identified as either milfoil (Myriophyllum sp.), coontail (Ceratophyllum demersum), algae, duckweed (Lemna sp.), Wolffia sp., sporangia (Chara sp.), or ‘other vegetation’, and seeds identified to genus. Food items were dried at 60o C for ≥ 48 hours before being weighed on a top-loading balance. Statistical Analysis To reduce the influence of rare occurrences when I encountered a duck that consumed a single food item in a large amount, I summarized diet data using a weighted, aggregate percent mass method explained in Swanson et al. 1974. I also divided the number of birds that consumed a particular food item by the number of birds in the sample to derive percent occurrence of food items. I summarized diet data for 2006 and 2007 for each individual duck species in 3 categories: vegetation, invertebrates, and seeds. For data to be used in a multivariate analysis of covariance (MANCOVA), I converted aggregate percent dry mass of food items found in the esophagus into proportions of invertebrates and seeds and used those values as dependant variables for 4 of the aforementioned species (mallard, blue-winged teal, scaup, and ring-necked duck). Because vegetation composed a large portion of gadwall diets, I included proportion of vegetation in diet as a third dependant variable for gadwall. To examine variability in diet composition among species during spring, I used a MANCOVA in which I included the effects of species (blue-winged teal, gadwall, mallard, lesser scaup, and ring-necked duck), study site (Cache River, Illinois River, Wisconsin, Scioto River, Lake Erie, and Saginaw Bay), collection date, habitat type (agricultural, seasonal emergent, permanent emergent, open-water, and bottomland hardwood), reproductive status (follicle development present or absent), and year (2006, 2007) (PROC GLM, MANOVA option; SAS Institute, Inc., Cary, NC). I conducted 2 MANCOVA’s for each species, 1 in which I considered the study site in which the duck was collected and 1 in which I considered the transect in which the duck was collected. Because a study site was not replicated in each of the transects, I had to consider them in separate MANCOVA’s. To determine if spring diet varied by latitude, longitude, or date within each species, I included the effects of study site or transect, date, habitat, reproductive status, and year; including date by site and site by year interactions as additional effects of interest (PROC GLM, MANOVA option; SAS Institute, Inc., Cary, NC). I only included reproductive status (i.e., hens that had entered rapid follicle development (RFD) vs. hens that had not) as an independent variable in MANCOVA models for blue-winged teal and mallards, because these were the only species I encountered that had started RFD. When evaluating diet of a particular species, I reduced initial MANCOVA models using a step-wise procedure by removing the nonsignificant interaction terms (P > 0.10 based on Type III sums of squares) to obtain a final reduced model that contained all main effects and significant interaction terms (Badzinski and Petrie 2006a). If an interaction term was significant, I conducted separate MANCOVA’s on year 1 and year 2 data to reduce confounding effects of interaction terms on main effects. Contrasts of the effects in the reduced MANCOVA model were adjusted using the Tukey-Kramer method (PROC GLM; SAS Institute, Inc., Cary, NC). Because large variances are typically associated with diet data, I considered data to be highly significant at P ≤ 0.05 or marginally significant at P ≤ 0.10 using the Type III sums of squares. RESULTS Summary Statistics We collected 919 ducks during the study; 402 in spring 2006 and 517 in spring 2007. Of these, 847 contained sufficient amounts of food to be included in analyses (n = 203 blue-winged teal, 188 mallards, 116 gadwalls, 135 lesser scaup, and 205 ring-necked ducks) (Table 1.1). Aggregate percent biomass estimates for invertebrates, seeds and vegetation consumed by each species are reported in Table 1.2. A more detailed description of diet at each study site is provided in Appendix A. Invertebrates that composed the largest portion of diet (i.e., greatest aggregate percent biomass) in dabbling ducks were: gastropods and Chironomidae in blue-winged teal, Chironomidae and microcrustacea (e.g., Cladocera, Copepoda and Ostracoda) in gadwall, and Chironomidae, macrocrustacea (e.g., Amphipoda and Isopoda), and non-dipteran insects in mallards (Table 1.3). The most common seeds were: Polygonum sp., Cyperus sp., Scirpus sp., and Leersia sp. in blue-winged teal, Polygonum sp., Cyperus sp. and Scirpus sp. in gadwall, and Polygonum sp., Leersia and Scirpus in mallards (Table 1.4). Lemna was the most commonly consumed vegetation by dabbling ducks, except gadwall, which consumed slightly more algae (Table 1.5). The most common invertebrates in diving duck diets were Chironomidae and Gastropods in scaup and Chironomidae and non-dipteran insects in ring-necked ducks (Table 1.6). The most common seeds found in scaup were Polygonum sp. and Potamogeton sp. and in ring-necked ducks were Polygonum sp., Potamogeton sp., and Echinochloa sp. (Table 1.7). Lemna and unidentifiable vegetation were the most Table 1.1. Number of ducks collected in the Upper MS River and Great Lakes Region that contained food items during spring 2006 and 2007. (BWTE = blue-winged teal, MALL = mallard, GADW = gadwall, RNDU = ring-necked duck, LESC = lesser scaup, CA = Cache River, IR = Illinois River, WI = Wisconsin, SR = Scioto River, LE = Lake Erie, SB = Saginaw Bay)
2006 2007 SPECIES CA IR WI SR LE SB Total CA IR WI SR LE SB Total TOTAL ________________________________________________________________________________________________ BWTE 22 21 19 2 20 12 96 27 28 17 10 6 19 107 203
________________________________________________________________________________________________ Table 1.2. Aggregate percent (A) ± standard error and percent occurrence (O) of food items in ducks (BWTE = blue-winged teal, MALL = mallard, GADW = gadwall, LESC = lesser scaup, RNDU = ring-necked duck) collected in the Upper MS River and Great Lakes Region during spring 2006 and 2007.
vegetation invertebrates seeds A O A O A O ______________________________________________________________________________________
Table 1.3. Aggregate percent biomass of animal foods consumed by dabbling ducks collected in the Upper MS River and Great Lakes Region during spring 2006 and 2007. If food items were < 1.0% aggregate mass, they were listed as trace (tr.). ______________________________________________________________________________________ Food Item BWTE (n = 155) GADW (n = 88) MALL (n = 92) Agg. % Agg. % Agg. % ______________________________________________________________________________________
Physidae 10.7 1.1 6.9 Planorbidae 15.5 tr. 3.5 Bivalvia 1.4 0.0 0.0 Sphaeriidae 1.4 0.0 0.0 Chironomidae 19.1 31.8 19.8 Non-Chironomidae Dipterans 6.6 14.4 11.7 Ceratopogonidae 2.7 4.1 2.2 Chaoboridae 0.0 1.1 0.0 Simuliidae 0.0 1.0 0.0 Stratiomyidae 2.1 2.8 2.7 Tabanidae tr. 1.0 1.1 Tipulidae tr. 1.1 2.2 Macrocrustacea 9.7 3.9 19.2 Amphipoda 4.5 2.3 6.9 Isopoda 5.1 1.6 12.2 Microcrustacea 7.2 16.3 2.1 Cladocera 3.9 5.4 tr. Copepoda 2.1 9.6 1.0 Ostracoda 1.1 1.2 tr. Annelida 5.0 4.2 9.4 Hirudinea tr. 0.0 1.4 Table 1.3 continued. ______________________________________________________________________________________ Food Item BWTE (n = 155) GADW (n = 88) MALL (n = 92) Agg. % Agg. % Agg. % ______________________________________________________________________________________
Table 1.4. Aggregate percent biomass of plant foods consumed by dabbling ducks collected in the Upper MS River and Great Lakes Region during spring 2006 and 2007. If food items were < 0.5% aggregate mass, they were listed as trace (tr.). ______________________________________________________________________________________ Food Item BWTE (n = 185) GADW (n = 88) MALL(n = 175) Agg. % Agg. % Agg. % ______________________________________________________________________________________
Agg. % Agg. % Agg. % ______________________________________________________________________________________
Tubers tr. tr. 9.3 ______________________________________________________________________________________ Table 1.5. Aggregate percent biomass of vegetation consumed by dabbling ducks collected in the Upper MS River and Great Lakes Region during spring 2006 and 2007. If food items were < 0.1% aggregate mass, they were listed as trace (tr.). ______________________________________________________________________________________ Food Item BWTE (n = 40) GADW (n = 82) MALL (n = 34) Agg. % Agg. % Agg. % ______________________________________________________________________________________
Table 1.6. Aggregate percent biomass of animal foods consumed by diving ducks collected in the Upper MS River and Great Lakes Region during spring 2006 and 2007. If food items were < 1.0% aggregate mass, they were listed as trace (tr.). ______________________________________________________________________________________ Food Item LESC (n = 113) RNDU (n = 109) Agg. % Agg. % ______________________________________________________________________________________
Agg. % Agg. % ______________________________________________________________________________________
Table 1.7. Aggregate percent biomass of plant foods consumed by diving ducks collected in the Upper MS River and Great Lakes Region during spring 2006 and 2007. If food items were < 0.5% aggregate mass, they were listed as trace (tr.). ______________________________________________________________________________________ Food Item LESC (n = 110) RNDU (n = 185) Agg. % Agg. % ______________________________________________________________________________________
Agg. % Agg. % ______________________________________________________________________________________
commonly consumed vegetation by diving ducks, although ring-necked ducks also consumed large amounts of Chara sp. (Table 1.8). MANCOVA results I applied an arcsine square-root transformation to the diet data; however this did not eliminate the nonnormality of the proportions in the diet. I therefore concluded that my data was robust to transformation. Additionally, because of a prevalence of zeros in my data, I decided against using compositional analyses to evaluate diet. Even though these assumptions were violated by analyzing my diet data in a MANCOVA, this approach, however, has been utilized in recent waterfowl diet studies and appears to be the most efficient method of evaluating waterfowl diet data (Afton et al. 1991, Badzinski and Petrie 2006). Breeding vs. Non-Breeding Thirty-three mallards and 9 blue-winged teal had entered RFD. Interestingly, the diets of both mallards (F1,13 = 0.64, P = 0.42 for invertebrates and F1,13 = 0.14, P = 0.71 for seeds) and blue-winged teal (F1,12 = 0.39, P = 0.53 for invertebrates and F1,12 = 0.41, P = 0.52 for seeds) in RFD were similar to diets of mallards and blue-winged teal not in RFD and when I excluded RFD females from analyses, it did not change results; therefore I did not consider them separately in subsequent analyses. Although not statistically significant, there were, however, higher mean proportions of invertebrates in diets of blue-winged teal (58% vs. 42%) and mallards (27.7% vs. 15.7%) that had begun RFD than those that had not begun RFD. Table 1.8. Aggregate percent biomass of vegetation consumed by diving ducks collected in the Upper MS River and Great Lakes Region during spring 2006 and 2007. ______________________________________________________________________________________ Food Item LESC (n = 31) RNDU (n = 49) Agg. % Agg. % ______________________________________________________________________________________
The final MANCOVA model (with non-significant interactions removed) evaluating duck diets considered only main effects of study site, species, date, habitat, reproductive status, and year. Significant effects of the model included study site, species, and date for both the proportions of invertebrates and seeds in the diet (Table 1.9). Blue-winged teal, gadwall, and scaup had similar proportions (P > 0.10) of invertebrates and seeds in their diets. Likewise, the diet of mallards was similar to ring-necked ducks (Figure 1.3). Invertebrate consumption significantly increased with date, whereas seed consumption decreased. Blue-winged Teal Diets Two-hundred and three blue-winged teal (n = 96 in 2006 and n = 107 in 2007) were included in the analysis evaluating diet at the scale of study site (Table 1.1). Likewise, 203 blue-winged teal (n = 69 from eastern transect and n = 134 from western transect) were included in the analysis evaluating diet at the scale of transect (n = 34 in 2006 and n = 35 in 2007 from the eastern transect and n = 62 in 2006 and n = 72 in 2007 from the western transect). In 2006, the first blue-winged teal was collected on 14 March and the last on 5 May and in 2007, the first was collected on 18 March and the last on 3 May (Table 1.10). The final reduced MANCOVA model evaluating latitudinal variation in diet (invertebrates and seeds) of blue-winged teal included only the main effects of site, date, habitat, reproductive status, and year. The percentage of invertebrates in the diet varied
Source DF Type III SS Mean Square F value Pr > F ________________________________________________________________________
Yüklə 0,59 Mb. Dostları ilə paylaş:
|