Waterfowl Foods and Use in Grain Sorghum and Other Managed Habitats

1

Waterfowl Foods and Use in Grain Sorghum and other Managed Habitats

in the Mississippi Alluvial Valley

Annual Report

Alicia J. Wiseman, Department of Wildlife and Fisheries, Box 9690, MississippiState

University, Mississippi State, MS39762, USA

Richard M. Kaminski, Department of Wildlife and Fisheries, Box 9690, MississippiState

University, Mississippi State, MS39762, USA

Erick J. Larson, Department of Plant and Soil Sciences, Box 9555, MississippiState

University, Mississippi State, MS39762, USA

Kenneth J. Reinecke, US Geological Survey, PatuxentWildlifeResearchCenter, 2524

South Frontage Road, Suite C, Vicksburg, MS39180, USA

Samuel K. Riffell, Department of Wildlife and Fisheries, Box 9690, MississippiState

University, Mississippi State, MS39762, USA

ABSTRACTGrain sorghum (milo) is an energy-rich food for waterfowl in the Mississippi Alluvial Valley (MAV) and other wintering grounds. Post-harvest manipulations of grain sorghum stubble may increase production of a second crop (i.e., ratoon), conserve waste grain (i.e., grain lost before or during harvest), and stimulate natural seed production. We sampled a total of 6 harvested grain sorghum fields at Cache River National Wildlife Refuge (NWR) in Arkansas and Grand Cote NWR in Louisiana during fall and winter 2006-2007 to evaluate (1) effects of post-harvest manipulations on abundance of ratoon grain sorghum, waste grain, and natural seeds; and (2) waterfowl use of managed croplands and natural wetlands. Our post-harvest treatments included mowing, crushing, and no manipulation (i.e., control) of grain sorghum stubble, each with and without nitrogen fertilizer. Preliminary analyses revealed greatest mean abundance of ratoon seed in non-fertilized control plots (64.42 kg [dry mass]/ha;

SE = 32.76) followed by fertilized control plots (31.16 kg/ha; SE = 15.47), non-fertilized crushed plots (20.20 kg/ha; SE = 12.67), fertilized crushed plots (10.46 kg/ha; SE = 7.86), fertilized mowed plots (4.18 kg/ha; SE = 3.28), and non-fertilized mowed plots (2.56 kg/ ha; SE = 2.56). Control and crushed plots conserved the greatest amount of waste grain (0.10 kg/ha; SE = 0.04 and 0.06, respectively). Non-fertilized crushed plots contained the greatest mean abundance of natural seeds (1.99 kg/ha; SE = 1.46). At Cache River NWR, we conducted 8 waterfowl surveys from December 2006 to February 2007; flooded grain sorghum attracted the greatest mean density of ducks (244 ducks/flooded ha/survey; SE = 137) whereas the greatest mean density of geese occurred in harvested rice fields (60 geese/flooded ha/survey; SE = 39). At Grand Cote NWR, we conducted 7 surveys; a flooded field of unharvested rice and barnyard grasses (Echinochloa spp.) attracted the greatest mean density of ducks (39 ducks/flooded ha/survey; SE = 9) and geese (3 geese/flooded ha/survey; SE = 3). We will repeat our study in winter 2007-2008 to increase sample size to a total of 22 experimental blocks distributed across the MAV in Arkansas, Louisiana, and Mississippi.

INTRODUCTION

The Lower Mississippi Alluvial Valley (MAV) in the Mississippi Flyway is an important region formigrating and wintering waterfowl (Reinecke et al. 1989). Originally,this region was a vast bottomland hardwood ecosystemcovering nearly 10 million ha (Fredrickson 2005,Reinecke et al. 1989). Today, only 2.6 million ha of forested arearemain (Twedt and Loesch 1999). Loss and fragmentation of habitat and riverine flood management haveinfluenced the capability of the MAV to support wintering waterfowl populations (Reinecke et al. 1988, Fredrickson 2005,Wilson et al. 2005).

Waste grain (i.e., seed lost before or during harvest) availablein agricultural fields has partially mitigated loss of historical foraging habitat for waterfowl in the MAV (Wright 1959, Delnicki and Reinecke 1986, Combs and Fredrickson 1996). However, recent research has revealed a decline in availability of waste rice and soybeans in the MAV (Manley et al. 2004, Kaminski et al. 2005, Stafford et al. 2006). Stafford et al. (2006) concluded thatestimates of waterfowl foraging carrying capacity by the Lower Mississippi Joint Venture (LMJV)were based on over estimates of waste agricultural seed availability.

Concurrent with decreased availability of waste agricultural seeds are burgeoning populations of lesser snow geese (Chen caerulescens) in North America(Baldassarre and Bolen 2006). The arctic breeding landscapes of these geeseare experiencing detrimental changes in plant communities (Abraham and Jefferies 1997). The MAV has attracted large wintering populations of lesser snow geese,and these birds may decreasefood resources for other waterfowl (Havens 2007). Therefore, management practices that conserve waterfowl food are being sought by wildlife managers.

Grain sorghum is an agricultural crop managed for waterfowl habitatin some areas of the United States(e.g., Illinois, southern Texas; Anderson 1959, Carter et al. 1989, Dennis 1996, Havera 1999). Where grown in the MAV, managers have observed little use flooded grain sorghum fields by snow geese. Researchers have observed snow geese in unflooded grain sorghum fields inTexas (e.g., Ballard 1993, Dennis 1997), but Glazener (1946) reported few or no snow geese in grain sorghum fields when rice or corn wasnearby.

After harvest, grain sorghum has the ability to regenerate from the root and stalk and produce a second seed head (i.e., ratoon). Ratoon crops are attainablein the MAV but generally not of sufficient quality or quantity to justify harvest (R. Dupree, U.S. Fish and Wildlife Service, personal communication). Unharvested ratoongrain sorghummay provide supplemental food for wintering waterfowl in the MAV. Waste grain sorghum andnatural seeds in managed fields also may providefood. Grain sorghum hasa metabolizable energy (ME)of3.49 kcal/g (dry mass) for blue-winged teal (Anas discors),comparable to the ME of corn for mallards (Anas platyrhynchos, 3.67 kcal/g; Reinecke et al. 1989, Sherfy et al. 2001, Kaminski et al. 2003). Natural seeds average ~2.5 kcal/g for mallards (Reinecke et al. 1989, Kaminski et al. 2003).

Considering the decreased availability of waste agricultural seeds for wintering waterfowl in the MAV and possible increased feeding competition from snow geese, development of management strategies to produce and conserve supplemental food for wintering waterfowl is warranted. Therefore, our objectives were to 1) estimate abundance of ratoon grain sorghum seed in response to post-harvest manipulations; 2) estimate abundance of waste grain sorghum in response to harvest; 3) estimate abundance of natural seeds in response to post-harvest manipulations; and 4) estimate density of waterfowl among winter flooded grain sorghum fields, other croplands, and moist-soil wetlands.

STUDY AREA

In 2006, we studied atCacheRiver and Grand Cote National Wildlife Refuges (NWRs). Cache River NWR is located26 km south of Augusta, Arkansas in the CacheRiver floodplain. It is 25,091 ha and has1,740 ha of cropland. Grand Cote NWR is located 9 km south ofMarksville, Louisiana. It is 2,459 ha with 825 ha of croplands and 336 haof managed moist-soil wetlands.

METHODS

Experimental Design

We established 1 experimental block in a representative portion of each available grain sorghum field at each study site. Each block contained 3 0.8-ha plots to which we randomly assigned 1 of 3 post-harvest treatments of the grain sorghum stubble: (1) mowing, (2) crushing, and (3) no treatment (i.e., control). We randomly assigned 1 of 3 post-harvest treatments to each plot, all treatments present in each block.. We applied treatments≤ 2 weeks after grain sorghum harvest and separated plots with a 30-m buffer of untreated harvested grain sorghum to ensure independence of applied treatments within blocks. We also randomly assigned a fertilization treatment to half of each plot (0.4 ha). We used fertilizer in the form of prilled (pelletized) ammonium nitrate. Prilled ammonium nitrate is not volatile as are urea-based nitrogen (N) sources and remains on soil until incorporated by precipitation. Refuge staff used a spin spreader to apply the fertilizer at a rate of~168kg/ha(~26kg N/ha; E. J. Larson, Mississippi State University Extension, personal communication).

Sampling Ratoon Grain Sorghum

We began data collection at both NWRs on 6 November 2006 after first frost. We useda 1-m2 sampling frame divided into 20 equal sized cells(25cm x 20cm). We randomly selected10 sampling siteswithin each half plot and selected the first sampling siteby walking a random number of steps (1-100) from arandom corner of eachhalf plot and placed the first sampling site within the experimental area. We selected the other 9 sampling sites by throwing the frame in a random direction and distance across the treatment area.

We considered any seed head a ratoon if it was attached to a previously cut stalk or a root sprout within the sampling frame. We clipped all ratoon seed heads that were rootedwithin the frame and placedthem in a labeled bag. Additionally, we randomly chose 3 cells within the sampling frame and measured(1)height (cm) of each ratoon plant within the cells, (2) whether theratoon growth originated froma stalk or root,(3) whether ratoon plants hada seed head oronly leaf growth, and (4) number seedlings germinated from waste grain.

SamplingWaste Grain and Natural Seeds

Wemeasuredabundance of natural seed and waste grain sorghum using a blower-vac method(Penny et al. 2006). We placed the blower-vac randomly within the1-m2sampling frame. We vacuumed for 10 seconds and placedcollected material in a labeled bag (Penny et al. 2006). If any portion of a waste grain sorghum head was within the blower-vac sampling tube, we clipped and collected that portion and combined it with the vacuumed material.

Penny et al. (2006) recommended use of the blower-vacon dry soils. Whenwetsoil conditions existed, we collectedwaste grain sorghum and natural seeds from within a plastic circular sampling frame (12.7-cm diameter, 4 cm in height) used with the blower-vac (Penny et al. 2006, Ripley and Perkins 1965. This approach ensured the same area was sampled under wet and dry conditions. We randomly placed the circular frame in the 1-m2 frame and pressed it 1 cm in the soil. Then, weextracted the wet litter and 1 cm of soil from the sample areausing a metal spatula. If a waste grain head was partially in the circular frame, we clipped and collected the portion within the frame. We placed each sample in a labeled plastic bag.

Sample Processing

We stored all samples in a freezer at -10˚ C until processed (Gray et al. 1999). After thawing frozen samples, we placed ratoon samples in paper bags and dried themto a constant mass at 87˚ C. We separated seeds from any plant material and re-dried the seeds to a constant mass (Gray et al. 1999). We weighed thesamples and recorded dry mass to the nearest 0.001 g.

To process vacuumed samples, we separated natural seeds and waste grain sorghumfrom litter using a series of sieves (Nos. 6 [3.35mm], 18 [1.00mm], and 50 [300 µm]) and forceps. We further separated natural seeds and grain sorghum, dried each component separately at 87˚ Cto a constant mass, and recordeddry masses of each to the nearest 0.001 g.

We processed waste grain and natural seed samples collected from wet sites using procedures similar for processing soil core samples(Manley et al. 2004, Stafford et al. 2005, Stafford et al. 2006, Kross et al. 2007). We thawed and soakedfrozen samples in a mixture containing a 3% solution of hydrogen peroxide, ≤ 250 cm3 of baking soda, and ≤ 1 L of water to oxidize clays, and then rinsed the samplesthrough the aforementioned sieves (Kross et al. 2007). We collected debris and seeds from sieves and air-dried both for 48 hours. Then, using a lighted1.25× magnifying lens, we removed and separated natural seeds and grain sorghum and driedand weighedeach to a constant mass as described above.

WaterfowlAbundance

Each study site hadwinter flooded grain sorghum, soybean, rice, and moist-soil wetlands. We surveyed waterfowl approximately weekly in available flooded croplands and moist-soil wetlands from December 2006 to February 2007. We conducted surveys in the afternoon because refuge regulations prohibit morning access to sanctuaries.

We conductedflush surveys followingroutine practices of biologists on the NWRs (Richard Crossett, Cache River NWR, personal communication). We recorded species specific abundance of all waterfowl flushed from each flooded cropland type or moist-soil wetland. We then determined the approximate area surveyed on aerial mapsandused ArcGIS 9.0 forcalculation ofsurveyed area and waterfowl density.

Preliminary Analyses

We calculated mean abundance(± SE) of ratoon grain sorghum and natural seedsfor each of 6 treatment combinations (i.e., 3 post-harvest stubble treatments, with and without fertilizer) across 6 study blocks (n = 6). We also calculated mean abundance (± SE) of waste grain for only the 3 mechanical treatments because waste grain abundance was not affected by fertilizer application. We computed the mean density (± SE) of ducks and geese (i.e.,birds/flooded ha/survey) across 7 surveys at Grand Cote NWR and 8 surveys at Cache River NWR.

RESULTS

The greatest mean abundance of ratoon seed was in non-fertilized control plots (64.42 kg [dry mass]/ha; SE = 32.76) followed by fertilized control plots (31.16 kg/ha; SE = 15.47), non-fertilized crushed plots (20.20 kg/ha; SE = 12.67), fertilized crushed plots(10.46 kg/ha; SE = 7.86), fertilized mowed plots (4.18 kg/ha; SE = 3.28), and non-fertilized mowed plots (2.56 kg/ ha; SE = 2.56; Fig. 1). Control and crushed treatments conserved the greatest abundance of waste grain (0.10 kg/ha; SE = 0.04 and 0.06, respectively) and mowed plots the least (0.04 kg/ha; SE = 0.01; Fig. 2). Mean abundance of natural seeds was greatest in non-fertilized crushed plots (1.99 kg/ha; SE = 1.46) followed by fertilized control plots (1.84 kg/ha; SE = 1.04), fertilized mowed plots (0.73; SE = 0.29), fertilized crushed plots (0.47 kg/ha; SE = 0.24),non-fertilized control plots (0.45 kg/ha; SE = 0.24), and non-fertilized mowed plots (0.20kg/ha; SE = 0.11; Fig. 3).

At Cache River NWR, harvested and unharvested flooded grain sorghum in the same fields attracted the greatest mean density of ducks (244 ducks/flooded ha/survey; SE = 137) followed by harvested soybean (47 ducks/flooded ha/survey; SE =23), moist-soil (45 ducks/flooded ha/survey; SE = 34), and harvested rice (21 ducks/flooded ha/survey; SE = 15; Fig. 4). Mean density of geese was greatest in harvested flooded rice (60 geese/flooded ha/survey; SE = 39) followed by harvested soybean (10 geese/flooded ha/survey; SE = 7) and harvested and unharvested grain sorghum in the same fields(2 geese/flooded ha/survey; SE = 2; Fig. 4). No geese were observed in moist-soil wetlands at either CacheRiver or Grand Cote NWRs. At Grand Cote NWR, a flooded field of unharvested rice and barnyardgrass (Echinochloa spp.) attracted the greatest mean density of ducks (39 ducks/flooded ha/survey; SE = 9) followed by moist-soil (30 ducks/flooded ha/survey; SE = 11), harvested soybean (29 ducks/flooded ha/survey; SE = 19), harvested rice (28 ducks/flooded ha/survey; SE = 10), and harvested and unharvested grain sorghum (14 ducks/flooded ha/survey; SE = 12; Fig. 5). Meandensity of geesealso was greatest in the flooded field of unharvested rice and barnyardgrass (3 geese/flooded ha/survey; SE = 3) followed by harvested rice (2 geese/flooded ha/survey; SE = 1) and harvested soybean (1 goose/flooded ha/survey; SE = 0.49; Fig. 5). No geese were observed in flooded harvested or unharvested grain sorghum at Grand Cote NWR.

FUTURE RESEARCH

In 2007-2008,we willrepeat our study at 6 sites in the MAV. Sites include Cache River NWR and Shadwick Farm (private land; PL) in Arkansas,York Woods (PL) in Mississippi, and Grand Cote and Lake Ophelia NWRs andDuck Creek (PL) in Louisiana. After combining data from both years, we will perform analysis on data from 22 experimental blocks (2006-2007,n = 6; 2007-2008,n = 16). Sites are grouped into northern (Cache River NWR, Shadwick Farm, and York Woods) and southern (Grand Cote and Lake Ophelia NWRs and Duck Creek) regions of the MAV possibly allowing for analysis of a latitudinal effect on ratoon grain production.

Wet soil conditions at Grand Cote NWR in 2006 forced us to use two different methods to collecting waste grain and natural seeds which may confound estimates of seed abundance with experimental treatments. Thus, in fall 2007 we will collect waste grain and natural seeds using both methods at the same sampling sites, compare seed abundance between collection methods, and calculate a correction factor.

We will subject data to rigorous statistical analyses during spring 2008. We will analyze ratoon and natural seed abundance data in a split-plot (“split” for fertilization treatment), randomized block design(RBD) using SAS PROC MIXED (SAS Institute 1999, Freund and Wilson 2003, Gutzwiller and Riffell 2007). We will use latitude at sites as a covariate and year as a random effect. We will analyze data on waste grain sorghum abundance using a complete RBD (without split plots) because waste grain sorghum abundance is not affected by fertilization. We may examine relationships between ratoon grain sorghum and natural seed abundance and fall temperature and precipitation. We will analyze waterfowl survey data from each study site separately using SAS PROC MIXED in a repeated measures analysis of variance (SAS Institute 1999, Freund and Wilson 2003, Gutzwiller and Riffell 2007).

ACKNOWLEDGEMENTS

Our project is supported by the U.S. Fish and Wildlife Service (FWS) Region 4 National Wildlife Refuge Division andMcIntyre-Stennisfunds to the Forest and WildlifeResearchCenter at MississippiStateUniversity, as well as in-kind support from U.S. Geological Survey,PatuxentWildlifeResearchCenter (Vicksburg, Mississippi).

We thankMessrs. James Kennedy, Paul Meng, and Stanley Shadwick for gracious use of their private lands as well as Bob Strader, Jonathan Windley, Mindy Gautreaux, Richard Crossett, Richard Dupree, and other FWS personnel for helping implement each aspect of this study on their respective refuges. We also thankH. Hagy, J. Callicutt, J. Mobley, C. Nelson, and A. Hannigan for field and laboratory assistance.

LITERATURE CITED

Abraham, K. F., and R. L. Jefferies. 1997. High goose populations: causes, impacts and

implications. Pages 7-22 inB. J. D. Batt, editor. Arctic ecosystems in peril: report of the Arctic Goose Habitat Working Group. Arctic Goose Joint Venture special publication, U.S. Fish and Wildlife Service, Washington, D.C., and Canadian Wildlife Service,Ottawa, Ontario.

Anderson, H. G. 1959. Food habits of migratory ducks in Illinois. Illinois History Survey

bulletin. 27: 289-344. Champaign, Illinois, USA.

Baldassarre, G. A., and E. G. Bolen. 2006. Waterfowl ecology and management, Second

edition. Krieger Publishing Company, Malabar, Florida, USA.

Ballard, B. M. 1993. Sorghum management for waterfowl wintering in southern Texas. Thesis,

TexasA & I University, Kingsville, USA.

Carter, P.R., D.R. Hicks, E.S. Oplinger, J.D. Doll, L.G. Bundy, R.T. Schuler, and B. J.

Holmes. 1989. Grain sorghum (milo). Alternative field crops manual. University of WisconsinExtension, Madison, Wisconsin,and CooperativeExtensionUniversity of Minnesota, St. Paul, Minnesota.

Combs, D. L., and L. H. Fredrickson. 1996. Foods used by male mallards wintering in

southeastern Missouri. Journal of Wildlife Management 60: 603-610.

Delnicki, D., and K. J. Reinecke. 1986. Mid-winter food use and body weights of mallards and

wood ducks in Mississippi. Journal of Wildlife Management 50: 43-51.

Dennis, M. H. 1996. Evaluation of sorghum stubble management alternatives in southern

Texas. Thesis, TexasA & MUniversity – Kingsville, USA.

Fredrickson, L. H. 2005. Contemporary bottomland hardwood systems: structure, function and

hydrologic condition resulting from two centuries of anthropogenic activities. Pages 19-35 in L. H. Fredrickson, S. L. King, and R. M. Kaminski, editors. Ecology and management of bottomland hardwood systems: the state of our understanding. University of Missouri-Columbia, Gaylord Memorial Laboratory Special Publication No. 10, Puxico, Missouri, USA.