2011 SOUTHEASTERN NATURALIST 10(4):687–702 Status and Distribution of Breeding Secretive Marshbirds in the Delta of Arkansas Michael J. Budd1,3 and David G. Krementz2,* Abstract - We surveyed the Lower Mississippi Alluvial Valley of Arkansas (“the Delta”) during the breeding seasons of 2005 and 2006 using the national marshbird monitoring protocol for secretive marshbirds. We detected and documented breeding by Podilymbus podiceps (Pied-billed Grebe), Ixobrychus exilis (Least Bittern), Rallus elegans (King Rail), and Gallinula chloropus (Common Moorhen). We detected but did not document breeding by Botaurus lentiginosus (American Bittern), Porphyrula martinica (Purple Gallinule), and Fulica americana (American Coot), all of which have been documented to breed in the Delta. Our estimated occupancy rates for breeding marshbirds in the study area ranged from a low of 6% for the King Rail in 2006 to a high of 27% for the Least Bittern in 2005. The range of these occupancy rates are low and reflect the rarity of secretive marshbirds in the Delta. Secretive marshbird occupancy rates were higher in the southern third of the Delta, probably because wetlands were more abundant or of higher quality there. Introduction Secretive marshbirds include species of the rail, bittern, and grebe families that primarily inhabit marshes, are dependent upon wetlands for their entire life cycle, and are difficult to detect. Their generally secretive nature, the endangered status of several races and populations, and continued loss of habitat warrant an examination of their status (Eddleman et al. 1988, Johnson et al. 2009, Ribic et al. 1999). In the Mississippi Alluvial Valley, the Podilymbus podiceps L. (Piedbilled Grebe) population is possibly decreasing (population trend score [PT] = 4; Patuxent Wildlife Research Center 2007), and severe deterioration in the future breeding sustainability is expected (threat to breeding score [TB] = 4, significant potential threat exists). Ixobrychus exilis Gmelin (Least Bittern) is widespread but declining (PT = 4), and there are significant potential threats to the breeding population (TB = 4). Additionally, the US Fish and Wildlife Service (USFWS) identified the Least Bittern as a species of special concern because they are relatively rare and we lack basic information on status and trends in most areas (USFWS 2002). Rallus elegans Audubon (King Rail) has exhibited a decrease in population size (PT = 5), and a severe deterioration in the future sustainability of breeding is expected (TB = 4). Many US states consider the King Rail threatened or of special concern for similar reasons. The population trends for Porphyrula martinica L. (Purple Gallinule) and Gallinula chloropus L. (Common Moorhen) 1Arkansas Cooperative Fish and Wildlife Research Unit, Department of Biological Sciences, 1 University of Arkansas, Fayetteville, AR 72701. 2US Geological Survey, Arkansas Cooperative Fish and Wildlife Research Unit, Department of Biological Sciences, 1 University of Arkansas, Fayetteville, AR 72701. 3Current address - 131 Greenbriar Circle, Crossett, AR 71635. *Corresponding author. Internet: Krementz@uark.edu. 688 Southeastern Naturalist Vol. 10, No. 4 are largely unknown (PT for both = 3), and some threat to breeding exists (TB = 4 and 3, respectively). The current status of secretive marshbirds in the Lower Mississippi Alluvial Valley in Arkansas (“the Delta”) is mostly unknown (K. Rowe, Arkansas Game and Fish Commission [AGFC], Humphrey, AR, pers. comm.). Meanley (1969) conducted extensive fieldwork on rails and other wetland-dependent birds in the Delta during the 1960s and described marshbirds in the vicinity of Stuttgart as “common”. The Breeding Bird Survey (BBS) has routes in the Delta that have been run for a number of years (Sauer et al. 2008), but two aspects of the BBS do not lend themselves well for surveying secretive marshbirds. First, secretive marshbirds are by nature difficult to detect, and often are only detected using callplayback methods, which are not part of the BBS protocol (Johnson et al. 2009). Second, the surveys rely on roads, which typically are not located in or near extensive marshes where these species occur (Johnson et al. 2009). Evidence of the difficulty in detecting secretive marshbirds is apparent in the Arkansas wetland BBS trend analyses for 1966–2006 as no Least Bitterns, Purple Gallinules, or Common Moorhens were detected. Additionally, only 2 Pied-billed Grebes and 2 King Rails were detected during that time period (Sauer et al. 2008). Within the Delta, recent breeding season records (AAS 2010) suggest that more secretive marshbirds are detected in the southern third of the Delta, possibly because wetlands in that region are more abundant or of higher quality. The Delta was once part of a vast wetland area comprised of mostly bottomland hardwoods as well as emergent and submergent wetlands, and prairie. Between the 1950s and the 1970s, this area suffered annual wetland losses exceeding 120,000 ha per year as a result of clearing and conversion to agriculture and aquaculture facilities (King and Keeland 1999, Wilen and Frayer 1990). It is unknown how this change in land-use has affected the use of these converted wetlands by secretive marshbirds, as well as their overall population status in this region. The objective of our research is to provide an update on the status and distribution of secretive marshbird populations that breed in the Delta. These estimates will provide a baseline for future marshbird population monitoring (Johnson et al. 2009). Methods Study area We conducted our study in the Delta, a region dominated by extensive agriculture with fragments of remnant bottomland hardwood forest (Fig.1; King et al. 2006). Wetland habitat types surveyed included bottomland hardwood stands, Taxodium distichum L. (Bald Cypress) bayous, Cephalanthus occidentalis L. (Common Buttonbush) swamps, Salix spp. (willow) swamps, Typha spp. (cattail) marshes, reservoirs with minimal vegetation, and wetlands with a mixture of habitat types. Sites included a mixture of public and private wetlands. Site selection We divided the Delta into 3 approximately equal regions (north, central, south) in which we then used stratified random sampling to select survey sites 2011 M.J. Budd and D.G. Krementz 689 (Fig. 1). The 3 strata we used included wetland area, marsh vs. swamp, and forest area adjacent to the wetland. Wetland area was based on the proportion of a 400-m radius circle, centered at where the observer stood, that was covered in water, with >50% coverage being a large wetland area (>2.5 ha), 10–50% being a medium-sized wetland area (0.5–2.5 ha), and <10% being a small wetland area (<0.5 ha). A marsh is a wetland characterized by herbaceous hydrophytic vegetation, and a swamp is defined as a wetland containing ≥30% woody vegetation (Cowardin et al. 1979). Adjacent forest area was based on the proportion of the 400-m radius circle that was covered in woody vegetation ≥6 m in height, with >50% coverage being a large amount of forested area, 10–50% being a moderate amount of forested area, and <10% being a minimal amount of forested area. We randomly selected 50 wetlands per region per year. We selected large wetland areas containing emergent vegetation non-randomly, as they only existed at national wildlife refuges (NWRs) and wetland reserve program (WRP) lands. We selected several sites haphazardly due to logistical problems or because Figure 1. The Delta of Arkansas (highlighted in gray) subdivided into north, central and south regions. 690 Southeastern Naturalist Vol. 10, No. 4 designated wetlands no longer existed. In those instances, we selected the wetlands closest to the randomly selected wetland. Marshbird surveys We conducted three 10-day survey rounds per region, with ≈20 days between each round of surveys from 16 April–8 July 2005 and from 3 April–21 June 2006, following the North American Marsh Bird Monitoring Protocol (Conway 2003). We used 2 observers in 2005 and 3 observers in 2006. We broadcast the breeding and territorial calls for the following species: Pied-billed Grebe, Least Bittern, King Rail, Rallus limicola Vieillot (Virginia Rail), Purple Gallinule, and Common Moorhen. We conducted surveys 30 min before sunrise to 2 hrs after, and from 2 hrs before sunset to 30 min after (Conway 2003). Surveys were not conducted during heavy rain, heavy fog, or wind speeds ≥19 km/hr. Sites were separated by ≥200 m to avoid double counting individuals (Conway 2003). We trained and tested observers on the calls of secretive marshbirds before surveying. In 2005, since we did not have information on detection probabilities for any of the marshbird species in the Delta, we attempted to make 15 visits to each site. According to Conway et al. (2004), conducting 15 visits would be the maximum needed to ensure >90% probability of detecting the species if it is present. During each 10-day period, we completed 5 surveys per site for a total of 15 surveys per site in 2005. Based on 2005 results, we determined we only needed to survey each site 9 times to obtain the same amount of uncertainty (MacKenzie and Royle 2005). Scaling back to 9 visits allowed us to increase the number of sites surveyed. In 2006, we conducted 4 visits during the first 10-day period, 3 during the second, and 2 during the third. We put more emphasis on surveying earlier in the season in case a drought occurred, as in 2005. Birds were only counted if found occupying the selected site. If birds were located on adjacent wetland units, or encountered during times of travel, these birds were recorded as opportunistic detections. We used the maximum number of individuals detected during any one visit at each site to determine the number of individuals counted overall for each marshbird species (Paracuellos and Telleria 2004). The total number of individuals counted overall is not an abundance estimate, but an estimate of the minimum number of individuals encountered. Occupancy estimation We used the program PRESENCE 2.0 to calculate the probability of occupancy (psi), and in a two-step process, we estimated the probability of detection (p) as a constant or as a function of observer. We used p for each species to estimate the probability of a false absence using the formula (1 - p)k, where k equals the average number of visits conducted (MacKenzie et al. 2006). The sample sizes in the analysis are smaller than the overall number of sites visited, as we eliminated sites with fewer than 5 visits in 2005 and fewer than 4 visits in 2006. We chose these cutoffs, as it was roughly half of the average number of visits per site for the respective year. We eliminated 4 sites for Pied-billed Grebes from occupancy estimates as those sites were predominately ditches which Pied-billed Grebes had not been documented using for breeding (Muller and Storer 1999). Occupancy 2011 M.J. Budd and D.G. Krementz 691 estimates for Common Moorhen and Purple Gallinule were not calculated since their sample sizes were too small to produce reliable estimates. In addition to testing for detection being a function of observer, we tested for a regional effect on occupancy across the Delta based on the observation that more secretive marshbirds occur in the southern third of the Delta (AAS 2010), as well as the null model. The regional test included all species in the analysis under the prediction that densities were highest in the southern region followed by the central and finally the northern region. We used Akaike’s information criterion (AIC) corrected for small sample size (AICc) to select among models. We considered models with ΔAICc values ≤2 to have strong support (Burnham and Anderson 2002). The detection and occupancy probability estimates from the null model are reported. Results We surveyed 190 sites (80 in 2005, 110 in 2006; Fig. 2). Overall, we surveyed 88 sites in the southern region, 61 in the central region, and 41 in the northern region. In 2005, 32 sites were on NWRs, 14 on private land, 15 on WRP lands, 16 on wildlife management areas (WMA), 2 on Army Corp of Engineer lands, and 1 was on National Park Service land. In 2006, 26 sites were on NWRs, 63 on private land, 16 on WRP lands, and 5 were co-managed as a WMA/WRP. In 2005, 21% of sites were surveyed 15 times (mean = 10.0). In 2005 the Delta experienced a drought, which resulted in 33 sites completely drying before surveying them 15 times. In 2006, 86% of sites were surveyed 9 times (mean = 8.45). In 2005, 54% of sites had ≥1 individual of a particular secretive marshbird species, and 28% of sites had >1 species. The average number of species per occupied wetland was 1.9 (SE = 0.16). The maximum number of marshbird species found at any one site was 5, which occurred at 1 site. In 2006, 46% of sites had ≥1 individual of a particular secretive marshbird species, and 56% of sites had >1 species. The average number of species per occupied wetland was 2.3 (SE = 0.21). The maximum number of marshbird species found at any one site was 6, which occurred at 1 site. In 2005, the top model selected was the null model with marshbird detection probability being 0.31 (SE = 0.02), and the occupancy probability being 0.45 (SE = 0.06). The next most plausible model included region, but the ΔAICc was 2.96. In 2006, the top model selected included a region effect, with the occupancy being highest in the southern region followed by the central region and finally the northern region. The 2006 null model detection probability was 0.44 (SE = 0.03), and the occupancy probability was 0.32 (SE = 0.05). Species accounts Pied-billed Grebe. In 2005, we detected Pied-billed Grebes at 9 sites and counted 28 individuals (mean count per site = 2.8, SE = 0.59). We opportunistically detected Pied-billed Grebes at 3 additional sites. In 2006, we detected Piedbilled Grebes at 20 sites and counted 53 individuals (mean = 2.5, SE = 0.44). We opportunistically detected Pied-billed Grebes at 10 additional sites. Of the 29 randomly selected sites occupied by Pied-billed Grebes in 2005 and 2006, 11 were on federal land, 7 on WRP land, 10 on private land, and 1 on a WMA (Fig. 3). 692 Southeastern Naturalist Vol. 10, No. 4 In 2005, a brood consisting of 3 young was detected at Cache River NWR, while in 2006, one active nest was detected at the Wallace Trust WRP site, in addition to a brood, consisting of 3 young, at a different location within the Wallace Trust WRP site. In 2005, 9 sites out of 69 were occupied, giving a naïve estimate of 0.13. The overall detection probably was 0.52 (SE = 0.05), and we found an observer effect. The occupancy probability was 0.13 (SE = 0.04). The average probability of a false absence was 0.0006. In 2006, 17 out of 88 sites were occupied, giving Figure 2. Distribution of sites surveyed for secretive marshbirds in 2005 and 2006 in the Delta of Arkansas. 2011 M.J. Budd and D.G. Krementz 693 a naïve estimate of 0.20. The overall detection probability estimate was 0.39 (SE = 0.04), while the occupancy probability was 0.21 (SE = 0.05). The probability of a false absence was 0.02. Least Bittern. In 2005, we detected Least Bitterns at 15 sites and counted 20 individuals (mean = 1, SE = 0.21). We opportunistically detected Least Bitterns at 2 additional sites. In 2006, we detected Least Bitterns at 20 sites and Figure 3. Distribution of Pied-billed Grebes using detections from randomly selected sites and opportunistic detections from 2005 and 2006 marshbird surveys in the Delta of Arkansas. 694 Southeastern Naturalist Vol. 10, No. 4 counted 37 individuals (mean = 1.9, SE = 0.25). We opportunistically detected the Least Bittern at 6 additional sites. Of the 25 randomly selected sites occupied by Least Bitterns for 2005 and 2006, 14 were on federal land, 8 on WRP land, 2 on private land, and 1 on a WMA (Fig. 4). Two active nests and two initiated nests were found on 6 June 2006 at the Wallace Trust WRP site. On 7 June 2006, 4 eggs were found in an active nest on the Chicot County WRP site, but this nest was found destroyed in late June. Figure 4. Distribution of Least Bitterns using detections from randomly selected sites and opportunistic detections from 2005 and 2006 marshbird surveys in the Delta of Arkansas. 2011 M.J. Budd and D.G. Krementz 695 In 2005, Least Bitterns occupied 15 sites giving a naïve estimate of 0.22. The overall detection probability was 0.16 (SE = 0.04), while the occupancy probability was 0.27 (SE = 0.06). The probability of a false absence was 0.17. In 2006, 16 out of 88 sites were occupied, giving a naïve estimate of 0.18. The overall detection probability was 0.58 (SE = 0.05), while the occupancy estimate was 0.18 (SE = 0.04). The probability of a false absence was 0.001. King Rail. In 2005, we detected the King Rail at 11 sites and counted 24 individuals (mean = 2.18, SE = 0.46). In 2006, we detected the King Rail at 6 sites and counted 18 individuals (mean = 3.0, SE = 0.68). We opportunistically detected the King Rail at 5 additional sites. Of the 17 randomly selected sites occupied by the King Rail, 6 were on federal land, 8 were on WRP land, 2 were on private, and 1 was managed as a WRP/WMA (Fig. 5). We observed one brood on 6 June 2006 at Hogwallow WRP. There were 5 young with 1 adult. The young were ≈60% of the adult’s size, making them about 30 days old (Meanley 1953). Backdating put the start of incubation at 15 April. In 2005, King Rails occupied 10 out of 69 sites, giving a naïve estimate of 0.14. The overall detection probability was 0.17 (SE = 0.04), and we found an observer effect. The occupancy probability was 0.22 (SE = 0.07). The probability of a false absence was 0.16. In 2006, 5 sites out of 88 were occupied, giving a naïve estimate of 0.057. The overall detection probability was 0.39 (SE = 0.08), while the occupancy estimate was 0.06 (SE = 0.03). The probability of a false absence was 0.02. Purple Gallinule. In 2005, we detected the Purple Gallinule at 2 sites and counted 3 individuals (mean = 1.5, SE = 0.5). We did not detect the Purple Gallinule in 2006, other than 1 opportunistic detection. We only detected the Purple Gallinule at 2 sites in the southern region in 2005 and 2006 (Fig. 6). Common Moorhen. In 2005, we detected the Common Moorhen at 2 sites and counted 6 individuals (mean = 2.8, SE = 2.0). We opportunistically located moorhens at 2 additional sites. In 2006, we detected the Common Moorhen at 4 sites and counted 11 individuals (mean = 2.8, SE = 0.48). Of the 6 randomly selected sites for 2005 and 2006, 2 were on federal land, 1 on WRP land, and 3 on private land (Fig. 7). Three nests were found at Wallace Trust WRP. Non-resident marshbirds In both years, we detected Botaurus lentiginosus Rackett (American Bittern), Virginia Rail, Porzana carolina L. (Sora), and Fulica americana Gmelin (American Coot) during our surveys, but we never recorded any evidence of breeding by these birds. The last date of detection in either year was 18 May 2006 for American Bitterns, 7 May 2006 for Virginia Rails, 27 May 2005 for Soras, and 18 May 2006 for American Coots. Discussion Survey considerations We found detection probabilities for secretive marshbirds ranged from a low of 16% for Least Bitterns in 2005 to a high of 58% for Least Bitterns in 2006. 696 Southeastern Naturalist Vol. 10, No. 4 Detection probabilities for the Pied-billed Grebe were consistently high each year (39–52%), probably because their loud calls carry well across open water where they occur; both Least Bittern and King Rail detection probabilities were lower in 2005 than in 2006. We suspect that the year effect was possibly a result of observer differences, as the observers used in 2006 had more experience with marshbirds than those used in 2005. The consequence of the different detection Figure 5. Distribution of King Rails using detections from randomly selected sites and opportunistic detections from 2005 and 2006 marshbird surveys in the Delta of Arkansas. 2011 M.J. Budd and D.G. Krementz 697 probabilities among species and within species among years resulted in markedly different probabilities of false absence. In general, the chances of not detecting a bird when it was in fact present drops to an acceptable level (<15%) when the number of survey visits to a site increase (MacKenzie et al. 2006). However, we found that even with an average of 10 visits per site as in 2005, for both Least Bittern and King Rail, the false absence probabilities were about 15% because of Figure 6. Distribution of Purple Gallinules using detections from randomly selected sites and opportunistic detections from 2005 and 2006 marshbird surveys in the Delta of Arkansas. 698 Southeastern Naturalist Vol. 10, No. 4 their low probabilities of detection. A 15% false absence probability may not be acceptable depending on the survey goals. We are aware of some marshbird surveys with usually only one visit per season (Dinsmore et al. 2010), even though Conway (2008) suggests a minimum number of 3 visits per site per season. If our detection probabilities for the Least Bittern and King Rail are representative for Figure 7. Distribution of Common Moorhens using detections from randomly selected sites and opportunistic detections from 2005 and 2006 marshbird surveys in the Delta of Arkansas. 2011 M.J. Budd and D.G. Krementz 699 marshbirds in general, then at 2 or 3 visits per season, the false absence probabilities will exceed 60%. Such false absence probabilities would not be acceptable, and not reporting these probabilities would be misleading to the reader. Status and distribution We confirmed the breeding of the Pied-billed Grebe, Least Bittern, King Rail, and Common Moorhen in Arkansas. Nests or broods were observed for all of these species. We observed a Purple Gallinule carrying nesting material at Arkansas Post National Park (NP), and AAS records report broods being observed at this site in previous years (AAS 2010). Breeding secretive marshbirds are uncommon across the Delta, with the Purple Gallinule and Common Moorhen being the least common breeding marshbirds there (James and Neal 1986). We detected Purple Gallinule at 2 sites: at Arkansas Post NP and the Wrape Plantation unit of Bayou Meto WMA. Only 1 other reported breeding site for the Purple Gallinule in the Delta (Desha County) has been reported (AAS 2010). Crow (1974) states that the Purple Gallinule is at the northern limits of its range, and likely expanded its range into Arkansas as rice farming expanded. We detected Common Moorhens at only 6 randomly selected sites overall. James and Neal (1986) report scattered nesting records in the Grand Prairie region of the Delta; however, we failed to detect the Common Moorhen in this region during our surveys. As early as the 1920s, the Pied-billed Grebe population in Arkansas was known to be in danger due to the drainage of wetlands and the demand for grebe feathers by milliners (Wheeler 1924). Howell (1911) reported the Pied-billed Grebe as being rare during the nesting season, with only one nesting observation in the southern region of the Delta. Wheeler (1924) reported that Pied-billed Grebes nested only in the Sunken Lands in the northeastern portion of the Delta, and at Wilmot in the southern region. The Arkansas Area Natural Plan indicates that Pied-billed Grebes have never been a common breeding species due to a lack of extensive permanent wetland complexes (Crow 1974), a notion also reiterated in James and Neal (1986). AAS (2010) records, Howell (1911), and Wheeler (1924) found Pied-billed Grebe more often in the southern and central regions of the Delta during the breeding season. We detected Least Bitterns at more sites than expected given their status as a species of concern. However, the overall occupancy estimates for both years were never >30%, reflecting that the species is still uncommon in the Delta. As recently as the 1960s, King Rails were considered common in the Stuttgart region of Arkansas (Meanley 1969). Rice farms dominate this region, and few natural wetlands exist there today. We surveyed along the same routes as Meanley (1953), as well as several rice fields in that area, but failed to detect the King Rail as well as either the Least Bittern or Purple Gallinule. Our surveys in this area were not part of the random sampling effort, but more of a focused effort to determine if King Rails could be found in this area. Changes in agricultural practices in the Stuttgart area include the dredging of irrigation ditches to keep them clear of vegetation, the planting of earlier maturing rice varieties, and the mowing of field edges (F. Lee, University of Arkansas, Rice Research and Extension Center, 700 Southeastern Naturalist Vol. 10, No. 4 Stuttgart, AR, pers. comm.). These changes have resulted in less cover available for nesting, and apparently have affected not only the King Rail, but also the Least Bittern and Purple Gallinule, which nested in irrigation ditches around Stuttgart (Meanley 1969). Three secretive marshbirds were primarily detected in the southern portion of the Delta: King Rail, Purple Gallinule, and Common Moorhen. When we tested across all secretive marshbirds for a regional effect by year, we found evidence for such an effect during both years, but only in 2006 was the model plausible. In 2005, the support for the region model was below our cutoff, but only by 1 ΔAICc. We suspect that the southern portion of the Delta had higher occupancy rates of secretive marshbirds because both the quantity and quality of wetlands found there were greater than in the two more northern regions. Of the marshbirds that we detected but did not find evidence of breeding for, only the American Bittern (James and Neal 1986) and American Coot (AAS 2010, James and Neal 1986) have been documented as breeding in the Delta. Breeding by both the American Bittern and American Coot is very rare, as both of these species normally breed farther to the north (Brisbin and Mowbray 2002, Gibbs et al. 1992). Our estimated occupancy rates for breeding marshbirds in the Delta ranged from a low of 6% for the King Rail in 2006 to a high of 27% for Least Bitterns in 2005. We were unable to estimate occupancy rates for both the Purple Gallinule and Common Moorhen. Even at the high end (27%), these occupancy rates reflect the rarity of these marshbirds in the Delta. Considering that Meanley (1969) described King Rails as common in the Delta during the 1960s, the King Rail in the Delta has clearly undergone declines over the past 50 years. We suspect that the primary mechanism for the decline in the King Rail, and probably for other secretive marshbirds in the Delta, is a consequence of loss of wetlands and reduction in quality of remaining wetlands (Dahl 2006, King and Keeland 1999, Wilen and Frayer 1990). Knowledge of the types of wetlands attractive to secretive marshbirds and how to manage those wetlands will be necessary to conserve secretive marshbirds in the Delta. While the types of wetlands and how to manage those wetlands for secretive marshbirds have been researched (Darrah and Krementz 2009, Rundle and Fredrickson 1981, Winstead and King 2006), Delta-specific wetland use and management guidelines are still needed (K. Rowe, pers. comm.). Acknowledgments Funding for this work came from the Arkansas Game and Fish Commission and the US Geological Survey Arkansas Cooperative Fish and Wildlife Research Unit. A. Claassen, J. Price, and S. Stake helped collect data, while L. Lewis and K. Rowe assisted with logistics. We thank all federal, state, and private landowners who allowed us to survey marshbirds on their properties. S. Lehnen assisted with the analyses. G. Huxel, R. Mc- New, J. Neal, and N. Winstead made comments on an early draft of this manuscript. Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the US Government. Two anonymous reviewers provided helpful comments on the manuscript. 2011 M.J. Budd and D.G. Krementz 701 Literature Cited Arkansas Audubon Society (AAS). 2010. Bird records database. Available online at http://www.arbirds.org/aas_dbase.html. Accessed 7 January 2010. Brisbin, I.L., Jr., and T.B. Mowbray. 2002. American Coot (Fulica americana). No. 697, In A. Poole and F. Gill (Eds.). The Birds of North America. The Birds of North America, Inc., Philadelphia, PA. Burnham, K.P., and D.R. Anderson. 2002. Model selection and multimodel inference: A practical information-theoretic approach. Springer-Verlag, New York, NY. 488 pp. Conway, C.J. 2003. Standardized North American marsh bird monitoring protocols. US Geological Survey, Arizona Cooperativ Fish and Wildlife Research Unit, Tucson, AZ. Conway, C.J., C. Sulzman, and B.E. Raulston. 2004. Factors affecting detection probability of California Black Rails. Journal of Wildlife Management. 68:360–370. Conway, C.J., C.P. Nadeau, R.J. Steidl, and A.R. Litt. 2008. Relative abundance, detection probability, and power to detect population trends of marsh birds in North America. Wildlife Research Report #2008-02. US Geological Survey, Arizona Cooperative Fish and Wildlife Research Unit, Tucson, AZ. 74 pp. Cowardin, L.M., V. Carter, F.C. Goley, and E.T. Laroe. 1979. Classification of wetlands and deepwater habitats of the United States. US Department of the Interior, Fish, and Wildlife Service, Washington, DC. FWS/OBS - 79/31. Crow, C.T. 1974. Arkansas natural area plan. Arkansas Department of Planning, Little Rock, AR. 248 pp. Dahl T.E. 2006. Status and trends of wetlands in the conterminous United States 1998 to 2004. US Fish and Wildlife Service, Washington, DC. Darrah, A.J., and D.G. Krementz. 2009. Distribution and habitat use of King Rails in the Illinois and Upper Mississippi River Valleys. Journal of Wildlife Management 73:1380–1386. Dinsmore, S.J, T. Harms, and D.H. Wilkens. 2010. Design and implementation of secretive marsh-bird monitoring. Available online at http://www.cfwru.iastate.edu/projects/ In%20Progress/Collaborating%20Faculty/contDinsmoreSecretMarshbird.pdf. Accessed 5 January 2011. Eddleman, W.R., F.L. Knopf, B. Meanley, F.A. Reid, and R. Zembal. 1988. Conservation of North American Rallids. Wilson Bulletin 100:458–475. Gibbs, J.P., S. Melvin, and F.A.Reed. 1992. American Bittern (Botaurus lentiginosus). No. 18, In A. Poole and F. Gill (Eds.). The Birds of North America. The Birds of North America, Inc., Philadelphia, PA. Howell, A.H. 1911. Birds of Arkansas. US Department of Agriculture, Washington, DC. Bulletin No. 38. 100 pp. James, D.A., and J.C. Neal. 1986. Arkansas Birds: Their Distribution and Abundance. University of Arkansas Press, Fayetteville, AR. 402 pp. Johnson, D.H., J.P. Gibbs, M. Herzog, S. Lor, N.D. Niemuth, C.A. Ribic, M. Seamans, T.L. Shaffer, W.G. Shriver, S.V. Stehman, and W.L. Thompson. 2009. A sampling design framework for monitoring secretive marshbirds. Waterbirds 32:203–215. King, S.L., and B.D. Keeland. 1999. Evaluation of reforestation in the lower Mississippi River alluvial valley. Restoration Ecology 7:348–359. King, S.L., D.J. Twedt, and R.R. Wilson. 2006. The role of the Wetland Reserve Program in conservation efforts in the Mississippi River Alluvial Valley. Wildlife Society Bulletin 34:914–920. Mackenzie, D.I., and A.J. Royle. 2005. Designing occupancy studies: General advice and allocating survey effort. Journal of Applied Ecology 42:1105–1114. 702 Southeastern Naturalist Vol. 10, No. 4 Mackenzie, D.I., J.D. Nichols, J.A. Royle, K.H. Pollock, L.L. Bailey, and J.E. Hines. 2006. Occupancy Estimation and Modeling: Inferring Patterns and Dynamics of Species Occurrence. Elsevier/Academic Press. Burlington, MA. 324 pp. Meanley, B. 1953. Nesting of the King Rail in the Arkansas rice fields. Auk 70:259–269. Meanley, B. 1969. Natural history of the King Rail. North America Fauna (67). Bureau of Sport Fisheries and Wildlife, Washington, DC. 108 pp. Muller, M.J., and R.W. Storer. 1999. Pied–billed Grebe (Podilymbus podiceps). No. 410, In A. Poole and F. Gill (Eds.). The Birds of North America. The Birds of North America, Inc., Philadelphia, PA. Paracuellos, M., and J.L. Telleria. 2004. Factors affecting the distribution of a waterbird community: The role of habitat configuration and bird abundance. Waterbirds 27:446–453 Patuxent Wildlife Research Center. 2007. Southeast US Waterbird Conservation Plan. Available online at http://www.pwrc.usgs.gov/nacwcp/southeast_us.html. Accessed 20 January 2011. Ribic, C.A., S. Lewis, S. Melvin, J. Bart, and B. Peterjohn (Eds.). 1999. Proceedings of the marsh bird monitoring workshop. US Fish and Wildlife Service Region 3 Administrative Report, Fort Snelling, MN. Rundle, W.D., and L.H. Fredrickson. 1981. Managing seasonally flooded impoundments for migrant rails and shorebirds. Wildlife Society Bulletin 9:80–87. Sauer, J.R., J.E. Hines, and J.Fallon. 2008. The North American breeding bird survey: Results and analysis 1966–2006. Version 10.13.2008. USGS Patuxent Wildlife Research Center, Laurel, MD. US Fish and Wildlife Service (USFWS). 2002. Birds of conservation concern 2002. Division of Migratory Bird Management, Arlington, VA. 99 pp. Available online at http:// fws.gov/pacific/migratorybirds/BCC2002.pdf. Accessed 15 January 2005. Wheeler, H.E. 1924. The birds of Arkansas: A preliminary report. State Bureau of Mines, Manufacturers, and Agriculture. Little Rock, AR. 177 pp. Wilen, B.O., and W.E. Frayer. 1990. Status and trends of US wetlands and deepwater habitats. Forest Ecology and Management. 33/34:181–192. Winstead, N.A., and S.L. King. 2006. Least Bittern distribution among structurally different vegetation types in managed wetlands of northwest Tennessee, USA. Wetlands 26:619–623.