REPORT OF ACTIVITIES 2017 TURTLE ECOLOGY RESEARCH REPORT Crescent Lake National Wildlife Refuge 31 May to 4 July 2017 A report submitted to Refuge Biologist Marlin French 15 July 2017 John B Iverson Dept. of Biology Earlham College Richmond, Indiana 47374 johni@earlham.edu PRIMARY PURPOSE: To continue studies of the natural history and population biology of the turtles on the Crescent Lake National Wildlife Refuge (now ongoing since 1981). METHODS: Our field techniques this year were basically the same as those used in previous years. We erected and monitored both our Main Gimlet Lake drift fence (900 meters, along the road paralleling the northeast lake margin) and Dike Road drift fence (200 m, parallel to the northwest margin of the lake) throughout the study period. Two Earlham undergraduate students and my wife Sheila assisted me with the field work this summer. CLIMATE BACKGROUND: March was warmer than normal, but this past spring was otherwise not atypical in terms of weather. The record high water table over these past two years had a much bigger impact on turtles than the spring weather. The drainage channel (under the road) from Crescent Lake was roaring when we arrived in late May but was only a shallow rivulet when we left in July. In May there were deep wetlands between our main fence and the sandhills to the east, which occurs only rarely. Thus, many mud turtles did not bother to travel all the way to Gimlet Lake when they emerged from hibernation on the sandhill, but rather just stayed in the new wetlands. Of course, this compromised our censusing at the fence. Hence, next year s spring censusing has become critical. However, June had near normal temperatures, but record low rainfall. We received only 0.25 of rain in June, compared to the 47-year mean of 2.78 (previous record lows were 0.52 in 1989 and 0.60 in 1994). RESULTS: Painted turtles (Chrysemys picta) During this year s field work, we processed 83 different female painted turtles (139 total captures) during their nesting forays. 56 of them were recaptures from previous years, but 29 were newly marked this year. We have now marked a total of 544 individual painted turtles from the Gimlet Lake wetland complex over the course of our 37 years of field work.
Nesting began shortly before our arrival on 31 May, so we unfortunately missed some early nests. Based on comments by Marlin French and temperatures in May, the estimated date of the first nest was about the 25th of May (long-term means are 28.5 May [N = 23 years data] and 7.5 June for average data of first nest [N = 19 years]. Our data analysis has shown that the timing of first nesting in painted turtles is inversely correlated with spring temperatures (warmer springs = earlier nesting seasons), but first nest timing has not advanced significantly over the past three decades despite global patterns of warming. Table 1. Yearly variation in dates of the first female painted turtle moving to nest, and the mean date for the production of the first nest by females in a given year (up to 3 or 4 clutches per year are possible by the same female). Year First female up to nest Mean date of females up to nest ---------------------------------------------------------------------------------------------- 1986 19 May 1.6 June 1988 23 May 29.9 May 1990 30 May 5.8 June 1993 22 May 4.3 June 1994 27 May ---- 1998 20 May ---- 1999 28 May ---- 2000 22 May 1.7 June 2001 25 May 1.5 June 2002 28 May 5.2 June 2003 29 May 8.1 June 2004 5 June 12.1 June 2005 15 June 25.4 June 2006 31 May 8.0 June 2007 27 May 6.2 June 2008 9 June 20.1 June 2009 6 June 16.4 June 2010 1 June 7.7 June 2011 6 June ---- 2012 9 May 19.2 May 2013 29 May 8.0 June 2014 27 May 7.1 June 2015 31 May 14.4 June 2017 ca. 25 May ca. 9 June Mean 28.5 May (N = 23) 7.4 June (N = 19 years)
The number of nesting female painted turtles this year was to that in 2015. We processed 92 nests this year (compared to 94 in 2015, 110 in 2013, and 75 in 2012). Instead of covering nests with protective hardware cloth screens, we left the nests in place to quantify natural depredation. As noted in other populations, depredation was time-dependent, most occuring in the first few days after deposition (Fig. 1). In addition, of the 19 nests deposited during the first week of June only eight (42.1%) were depredated by the time we left on 5 July. Depredation rate for the next week in June was similar at 6 of 19 nests (42.9%). However, for the next three weeks 16 of 19 (84.2%), 17 of 22 (77.3%), and 10 of 15 (66.7%) nests were depredated by 5 July. Clearly nests laid early in the season have the highest chance of surviving depredation, presumably because the main predators (raccoons and badgers) had not yet discovered the resource and shifted their search strategy to include nightly forays in the nesting areas. Fig. 1. Depredation of painted turtle nests following deposition. Most nests are destroyed within the first three days, but some are still taken weeks later. As usual, we mapped all nests for further monitoring. We also began work on precisely geomapping all nests and landmarks in the painted turtle nesting areas. Earlham Professor Dr. Jose Ignacio Pareja visited our field site in mid-june along with a cm-accurate GPS Trimble unit. We georeferenced as many landmarks and nests as we could during his three-day visit. Once all the nest sites from all the years are digitized we can examine spatial patterns of nesting across clutches, females, and years. We hope to make progress on this during the fall semester. Our future plan (assuming Refuge permission) is to return in late March or early April next spring to excavate the remaining painted turtle nests to determine nest survival. We want to test the hypothesis that nesting close of other nests confers a survival advantage to those nests. Predators will certainly destroy many nests laid in close proximity, but more may be overlooked than if they were laid farther apart. All live hatchlings from the surviving nests will be released immediately into Gimlet Lake.
One of research goals this coming year will be to analyze all our painted turtle reproductive data to look for effects of annual climate variation. We are curious to see if wetter or drier years or colder or warmer springs or autumns impact clutch size or egg size. Snapping turtles (Chelydra serpentina) The timing of nesting in snapping turtles this year (beginning 4 June and averaging 9 June) was about average (6 June and 12 June; N = 19 years; Table 2). We monitored 24 nesting females at Gimlet Lake this year (18 of them previously marked) and we have now marked a total of 104 snapping turtles (mostly females) in our Gimlet Lake study area since 1981. We again noted that the number of nesting snappers at Island Lake continues to be much reduced compared to early years. Table 2. Variation in nesting season and reproductive output of Nebraska snapping turtles at the CLNWR. Year Nesting season (n) Mean date Clutch size range (n) Clutch size mean ----------------------------------------------------------------------------------------------------- 1990 11-28 June (7) 18 June 37-73 (6) 54.8 1993 12-28 June (36) 20 June 29-71 (23) 48.3 1994 1-13 June (32) 6 June 31-72 (28) 45.4 1998 6-23 June (14) 11 June ------ --- 1999 4-22 June (29) 8 June 9-82 (27) 51.5 2000 29 May-12 June (21) 7 June 37-76 (19) 55.9 2004 29 May-15 June (15) 6 June 35-71 (12) 56.7 2005 7-21 June (36) 14 June 12-76 (33) 49.6 2006 1-11 June (29) 4 June 26-66 (24) 43.2 2007 28 May-14 June (91) 7 June 22-71 (32) 43.7 2008 15-25 June (54) 20 June 20-79 (44) 44.4 2009 11-24 June (65) 18 June 5-71 (39) 45.9 2010 10-27 June (52) 19 June 19-64 (39) 41.4 2011 14-25 June (21) 20 June 33-61 (17) 48.9 2012 1-8 June (57) 4 June 20-91 (49) 49.8 2013 7-17 June (49) 12 June 26-78 (35) 48.8 2014 3-23 June (29) 13 June 27-62 (26) 46.9 2015 7-14 June (45) 10 June 33-73 (38) 48.3 2017 4-20 June (43) 9 June 23-61 (29) 47.8 Means 5.6-19.5 June (19 yrs) 12 June 5-91 (520) 47.5 Although we have previously documented a significant inverse correlation between spring daytime temperatures and nesting season for snappers (see previous reports), there has been no pattern of earlier nesting over the course of the last three decades. This is surprising given the pattern of global warming observed elsewhere in North America. We have been analyzing past climate effects on the reproductive output (body-size-standardized clutch size, egg size and clutch mass) in snappers based on our long-term data. We discovered that warmer temperatures in September and October result in larger clutches, eggs, and clutch masses the following spring, but that spring-time temperatures did not affect these variables. However, the effects differed across female body size. For smaller females (less than 28 cm
carapace length) warmer autumns resulted in larger clutches but not egg sizes the next spring, but for larger females (greater than 28 cm CL) warmer autumns resulted in increased egg size but not clutch size. These results suggest that climate change could greatly affect reproduction in snapping turtles in the future. Yellow mud turtles (Kinosternon flavescens) As noted above the water table on the Refuge was at record height this year. As a result, many mud turtles that emerged from hibernation in the sandhills northeast of Gimlet stopped their normal migrations short and stayed in the temporary wetlands between the hills and our fence. As those wetlands dried down during our record dry June, many (most?) of those mud turtles moved west toward the permanent water in Gimlet. During that movement they were finally intercepted at our main fence. Thus, many juveniles and females did not reach our fence until mid or late June when they would normally reach the fence in late April or May. We were still capturing young of the year at the fence as late as 30 June that had spent time in these wetlands and grown significantly before heading for Gimlet. Unfortunately, this greatly complicated our population size assessments this year. Table 3. Yearly variation in dates of the first female mud turtle moving to nest, and the median date for all females nesting in a given year. Number of gravid females in parentheses. Main fence operated from 1981 to 1985, but extended around the Willows in 1986. Dike fence added in 1990. *Dike fence only. Year First female up to nest Median date of females up to nest ---------------------------------------------------------------------------------------------- 1981 10 June 19 June (126) 1982 7 June 27 June (55) 1986 29 May 19 June (204) 1988 29 May 10 June (208) 1990 9 June 23 June (324) 1993 1 June 23 June (361) 1994 28 May 9 June (406) 1998 26 May 14 June (352) 1999 9 June 21 June (343) 2000 1 June 10 June (413) 2004 27 May 20 June (285) 2005 9 June 24 June (241) 2006 26 May 11 June (254) 2007 4 June ---- 2010 7 June 19 June (336) 2013 2 June 16 June (145)* 2014 1 June 13 June (388) 2015 30 May 16 June (254) 2017 31 May 15 June (289) Mean 2.0 June (N = 19) 17.2 June (N = 18) The first gravid females (as confirmed by x-ray) began moving up to nest on 31 May (compared to the long-term average date of 2 June; Table 3). Median date this year for females moving up to the sandhills to nest (15 June) was also very near the long-term average (17 June). All
females heading up to nest (N = 289; Table 3) were xrayed to determine clutch and egg sizes. It is not yet clear what climatic factors drive some female mud turtles to skip reproduction in a given year, but we suspect that key drivers might be cold temperatures in autumn, cold temperatures in May and drought in the preceding year. We are currently trying to tease apart this story quantitatively; however, depredation is a compounding factor. For example, along the southern portion of the Main Fence (around the Willows wetland), there was heavy depredation on mud turtles in 2015 and this year by the many raccoons living in the willow trees (e.g., we found nine fresh raccoon-depredated shells of female mud turtles on the hillside above the wetland in June 2014, and seven in June 2017). This unusually high depredation complicated our ability to estimate reproductive frequency in the mud turtle. Since the study began in 1981, we have individually marked a total of 5119 mud turtles (25 new this year, plus another 26 that were cohort-marked as hatchlings and were given individual marks this year) and we have tabulated (at least) 26747 total captures, although we did not have time to mark any juveniles this year (despite nearly 400 juvenile captures). Our long-term clutch histories, resulting from our xraying of all gravid females since 1988, are being related to local climatic variables to examine which of the latter correlate with higher or lower reproductive output in a given year. This analysis will be completed this fall. Finally, in an analysis of the timing of nesting in mud turtles, as for painted and snapping turtles, we found a significant inverse correlation between spring day-time temperatures and nest dates, but no consistent pattern of change over the the last three decades. Hence, although climate change has been significant on the Refuge over the past 37 years, it has not so far affected average nesting timing for any of the above three turtle species. Ornate box turtles (Terrapene ornata) This year we captured only 31 individual box turtles. Of these, 19 were recaptures from previous years and 12 were newly marked. We have now marked a total of 597 box turtles (at least 2640 total captures) in the Gimlet Lake study area. The number of box turtle captures was very low this year, presumably because of the low June rainfall (rain stimulates their activity). We did recapture two female box turtles first marked in 1981. One (#2) had grown 7 mm in the intervening 36 years (114 to 121 mm plastron length), and the other (#8) had not grown at all (118 mm PL) over that time. A manuscript summarizing all of our 37 years of reproductive and growth data for this species is nearly complete. Hognose snakes (Heterodon nasicus) and Bullsnakes (Pituophis catenifer) We captured three gravid bullsnakes and two gravid hognoses during June, and held them until they laid egg clutches. We then took a tissue sample from each female, and released the females where they were collected. We brought the eggs back to Earlham to incubate to hatching, after which we intend to take tissue samples from each hatchling for paternity analyses. We will overwinter the hatchlings at Earlham and release them where the females were captured early next April. Our eventual goal is to accumulate a large enough sample to do a paternity analysis for both hognose snakes and bullsnakes.
ACKNOWLEDGMENTS: The support and collegiality of the Refuge staff is always appreciated. We especially value the vigilance of Marlin French in spotting painted turtles on nesting forays that we might have missed. FUTURE PLANS: Tentative: 2018 - The last time I monitored the mud turtle emergence from hibernation was 2007, and that element of my work is long overdue. I would like to arrive at the Refuge on ca. 1 April and stay until the end of May, erecting and monitoring both turtle fences for that period. My wife and I would like to be here for the month of April, and be joined by one or two students during May. 2019: - Because of the uncertainty of funding for 2019, our plans for 2019 and beyond are still uncertain.