Effects of Carp on the Survival and Growth of Wild Rice in ... · Final report to St. Croix Tribal Environmental Services – Natural Resources Department, Webster (WI). Freshwater

1

Effects of Carp on Wild Rice Growth in Upper Clam Lake – November 2010Freshwater Scientific Services, LLC

Effects of Carp on the Survival and Growth of Wild Ricein Upper Clam Lake – Burnett County, WI

Prepared for St. Croix Tribal Environmental Services – Natural Resources Department, November 2010by James A. Johnson – Freshwater Scientific Services, LLC

www.fixmylake.com18029 83rd Avenue NorthMaple Grove, MN [email protected](651) 336-8696

1


Financial and technical support for this project was provided by the following organizations andindividuals:

St. Croix Tribal Environmental Services – Natural Resources Department! Anthony J. Havranek *! Jamie Thompson! Chad Songetay! Dave Spafford

Freshwater Scientific Services, LLC! James A. Johnson *

Great Lakes Indian Fish & Wildlife Commission! Peter David – Wildlife Biologist

Bureau of Indian Affairs

Clam Lake Protection and Rehabilitation District

Wisconsin Department of Natural Resources

* Principal investigators

Cite as:Johnson JA, Havranek AJ. 2010. Effects of carp on the survival and growth of wild rice in Upper Clam Lake – BurnettCounty, WI. Final report to St. Croix Tribal Environmental Services – Natural Resources Department, Webster (WI).Freshwater Scientific Services LLC, Maple Grove (MN).

2


IntroductionOver the past 20 years, Upper Clam Lake (Burnett Co., WI) has consistently supported extensive stands (250 to300 acres) of northern wild rice (Zizania palustris L.). Beginning in 2007, the lake experienced a dramaticdecline in both the extent and density of wild rice growth. This decline continued through 2010, withadditional decreases in the extent and density of rice stands. As of 2010, rice growth in Upper Clam Lake waslimited to sporadic individual plants found immediately adjacent to shore, with a few sparse remnant stands inisolated shallow bays. Previous studies have reported that wild rice beds typically experience natural declinesin about one out of every four years (Moyle 1944; Walker et al. 2010). However, even after substantial naturaldeclines, rice beds typically reestablish quickly from seeds remaining in lake sediments (Moyle 1944). In UpperClam Lake, the decline of wild rice that began in 2007 was particularly severe and persisted for fourconsecutive years. Furthermore, a seed enumeration study conducted in 2009 and 2010 (Johnson 2010)showed that very few wild rice seeds remained in the sediments of Upper Clam Lake. These observationssuggest that the decline of rice in the lake was not a typical natural decline.

The decline of wild rice in other lakes has been associated with water level fluctuations (Moyle 1944),unfavorable weather (Moyle 1944), low nitrogen availability in sediments (Walker et al. 2010), and destructionof rice plants by carp and muskrats (Moyle 1944). Although several of these factors may have contributed tothe decline of rice in Upper Clam Lake, local residents and natural resources managers suggested that thedecline coincided with an increase in the population of common carp (Cyprinus carpio L.). Previous studieshave reported that carp can severely reduce aquatic vegetation by uprooting plants (Crivelli 1983; Sidorkewicj1998), eating plant shoots (King and Hunt 1967; Sidorkewicj 1998), feeding on seeds in lake sediments (Crivelli1983), or by increasing turbidity (Moyle and Kuehn 1964; Sidorkewicj 1998). Furthermore, Moyle (1944)reported that many wild rice stands in southern Minnesota disappeared after carp invaded area lakes in theearly 1900s. The Wisconsin Department of Natural Resources (WDNR) and St. Croix Tribal EnvironmentalServices – Natural Resources Department conducted a survey in 2009 (electrofishing in shallow bays) to assessthe age-structure of the carp population in Upper Clam Lake. Results from this survey indicated that carp had avery successful hatch in 2005, with high survival in subsequent years. Given the documented heavyrecruitment of carp in 2005 and its coincidence with the decline of rice in the lake, resource managersconcluded that carp likely played a role in the decline of rice. However, prior to 2010, managers were unable todetermine the impact of the carp relative to other factors known to affect wild rice survival and growth (e.g.water level, water clarity, weather, and disease; Moyle 1944).

Although Moyle (1944) anecdotally implicated carp for the destruction of wild rice stands in southernMinnesota, we found no studies that specifically evaluated the effect of carp on wild rice. Furthermore, little isknown about the persistence of wild rice seed in lake sediments over multiple years or the minimum seedabundance needed for natural recovery of rice stands after major declines. This study was designed to assess(1) the importance of direct carp effects (herbivory and uprooting) on wild rice survival, growth, andreproduction in Upper Clam Lake, (2) whether sufficient seed remained in the lake’s sediment in 2010 to allowfor rapid natural recovery in the absence of carp, and (3) whether manual seeding could enhance rice survivaland growth in the lake. The results presented here will help to guide management in Upper Clam Lake byevaluating whether a reduction in carp abundance and/or manual seeding will be needed to reestablish wildrice stands in the lake. Moreover, this study advances knowledge of the effects that carp have on wild rice, andmay help to improve future management of rice stands in other areas.

3


Study LakeUpper Clam Lake (45°47’25”N, 92°19”30”W; WBIC 2656200) is alarge (1207 acres) shallow, eutrophic drainage lake (11 ft maxdepth, 5 ft mean depth) in Burnett County, WI (Figure 1). Thelake’s hydrology is largely driven by the Clam River, which drainsan area of approximately 250 square miles upstream of the lake.Upper Clam Lake is connected to Long Lake (southwest) by asmall intermittent stream, and to Lower Clam Lake (immediatelydownstream) and the Clam River Flowage (15 miles downstream)by the Clam River. Water levels in Upper Clam Lake are controlledby a dam on the Clam River downstream (north) of Lower ClamLake. The water level in the lake is typically drawn down by about1 ft in the fall of each year to reduce ice heave on shorelines andprovide additional water storage capacity during subsequent highspring inflow. Summer total phosphorus (96 ± 14 µg/L; mean ±SD) and chlorophyll-a (30 ± 5 µg/L) are quite high in Upper ClamLake (2009 and 2010 data). Consequently, summer water clarity istypically low (Secchi depth = 2.6 ± 0.1 feet).

Figure 1. Location of Upper Clam Lake

Figure 2. Map showinglocation of Upper ClamLake, Lower Clam Lake,Long Lake, Clam RiverFlowage, and the ClamRiver (flow from SE to NW).

1 mile

N!

Clam RiverFlowage

Lower ClamLake

Upper ClamLake

Long Lake

Clam River

4


History of Wild Rice in Upper Clam LakeWild rice harvest records (Figure 3) show that Upper Clam Lake was a consistent source of wild rice from 1992through 2006. Similarly, photographs taken prior to 2007 showed that the lake had supported extensive, lushstands of wild rice as recently as 2005 and 2006 (Figure 4a,b). However, after 2006, there was no reportedharvest of rice from the lake, and aerial images taken in June 2008 showed that very little rice remained inareas that had historically supported lush rice stands (Figure 4c). Furthermore, delineations of rice beds inUpper Clam Lake (boat surveys), conducted by St. Croix Tribal Environmental Services – Natural ResourcesDepartment in 2001 and then annually from 2007 through 2010, documented a dramatic decline in the extentand density of rice beds. These surveys showed that rice beds decreased from nearly 300 acres of lush ricegrowth in 2001, to about 60 acres of very sparse growth in 2010, mostly confined to shallow bays and sporadicpatches immediately adjacent to shore.

Figure 3. Annual wild rice harvest(lbs) for Upper Clam Lake, Long Lake,and nearby Briggs Lake: 1992-2009.Data provided by Great Lakes IndianFish & Wildlife Commission (GLIFWC).

Figure 4. Historical images of wild rice stands inUpper Clam Lake, WI; (a) large stand of rice in 2006(southwest shore of Lonestar Bay); (b,c) aerialimages showing extent of wild rice growth inUpper Clam Lake before (b; June 3, 2005) and after(c; June 23, 2008) the decline of rice stands in thelake (aerial images: Google Earth, accessed Nov2010).

a

WildRice

June 2005

June 2008

June 2005

June 2008

bb

cc

LonestarBay

Wild Rice Harvested: 1992-2009

0

2000

4000

6000

8000

10000

12000

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

Poun

ds H

arve

sted

Upper Clam LakeLong LakeBriggs Lake

Annual Wild Rice Harvest: 1992 to 2009

5


Wild Rice Seeds in Upper Clam LakeA wild rice seed enumeration survey was conducted on Upper Clam Lake and two neighboring lakes (LongLake and the Clam River Flowage) in the fall of 2009 and early spring of 2010 (Johnson 2010). Results fromthese surveys indicated that there were very few wild rice seeds remaining in the top 10 cm (! 4 in) ofsediment in Upper Clam Lake. Furthermore, these surveys showed that the abundance of seeds in the top 10cm of Upper Clam Lake (0 seeds m2) was significantly lower (P < 0.05, Welch’s t-test) than in Long Lake (250 ±43 seeds/m2; mean ± SE) or the Clam River Flowage (80 ± 32 seeds/m2). This suggests that the bank of wildrice seed in Upper Clam Lake had been severely depleted, making a natural recovery of rice stands in the lakeunlikely. A more detailed account of the wild rice seed enumeration methods and results is presented in theWild Rice Seed Enumeration Report: 2009-2010 (Johnson 2010).

Carp in Upper Clam LakeAnecdotal reports from local residents and natural resources managers suggested that the carp population inUpper Clam Lake increased noticeably between 2000 and 2010 while the panfish population had declinedover the same period. In June 2009, staff from the Wisconsin DNR (Spooner, WI) and St. Croix TribalEnvironmental Services – Natural Resources Department conducted electrofishing in shallow near-shore areasof Upper Clam Lake. During this survey, they captured and measured approximately 300 carp, with subsequentage determination (spinal annuli) for 140 of these captured carp. Results indicated that over 40% of the carpwere from a single year-class that had hatched in 2005 (Figure 5). This suggests that conditions in the springand summer of 2005 favored carp recruitment. Furthermore, by 2009, carp in this abundant year-class hadgrown large, generally ranging from 20 to 25 in long, suggesting that the population would experience little orno loss due to predation in subsequent years. The documented abundance and survival of these large carp inUpper Clam Lake suggested that there was a high potential for impacts on aquatic plants, particularly duringspring spawning when the carp concentrated in shallow areas. A 2 year-old gravid female carp was capturedduring the 2009 survey, suggesting that the large group of carp born in 2005 may have begun spawning inshallow areas as early as 2007, when they reached 2 years of age. This timing coincides with the first year ofmajor wild rice decline in the lake.

Figure 5. 2009 carp population agestructure in Upper Clam Lake, WI. Age dataprovided by the Wisconsin DNR and St.Croix Tribal Environmental Services –Natural Resources Department. Fish agewas determined by evaluating dorsal spineannuli from 140 captured individuals.

Other Ages13%

Age 223%

Age 204%

Age 199%

Age 185%

Age 174%

Age 169%

Age 154%

Age 142%

Age 135%

Age 442%

6


MethodsStudy SitesIn early April 2010, we selected three study sites in shallow, near-shore areas of Upper Clam Lake that hadsupported stands of wild rice in years prior to 2007 (Figure 6a). At each of these sites, we identified a 150-fttransect oriented parallel to the nearest shoreline, with uniform water depth and sediment texture (Table 1).Along each of these transects, we randomly placed 2 open plots (12 " 12-ft; no fence) and 2 fenced plots (12 "12-ft), maintaining a distance of approximately 25 feet between plots.

All open and fenced plots were installed on April 5, 2010, approximately 1 week after ice-out. Open sites weredelineated by marking the corners of each plot with posts (Figure 6c). Fenced plots (carp exclosures) wereconstructed with 5-ft high, 14-gauge, galvanized welded-wire fence (1 " 2-in mesh), anchored at each cornerand at the midpoint of each side with PVC pipe driven 5 to 6 ft into the sediment (Figure 6b). Prior to installingthe exclosures, we attached 2.5-in diam " 5-ft PVC sleeves to the fence, spaced 6 ft apart. These sleeves madeit easier to position and anchor the exclosures in the field. During installation, the fence was held in positionwhile smaller diameter PVC pipes (1.5-in diam " 10-ft) were pushed through the sleeves and into thesediment. Once all corners and sides of each exclosure were anchored, the fence was pushed 12 to 18 inchesinto the sediment to prevent carp from burrowing under the fence. At deeper sites (A and B), we addedvertical fence extensions (18 to 20 inches high) to prevent carp from jumping into the exclosures. These fenceextensions were constructed with galvanized poultry fencing (24-in high; 1-in mesh) which was attached toadditional PVC anchor posts driven into the sediment (Figure 6b). The bottom of the extension fence waspositioned to overlap the outside of the heavier exclosure fence by several inches, and the two fence layerswere tied together with plastic cable ties to prevent billowing and gaps. We inspected the fenced exclosuresroughly every two weeks to assess any need for repairs and removed any floating debris that had accumulatedon the fence.

Manual Seeding of PlotsIn the late summer of 2009, St. Croix Tribal Environmental Services collected wild rice seed from standing riceplants in Long Lake (immediately southwest of Upper Clam Lake) to use for manual seeding. The collectedseed was stored in large, mesh, nylon bags which were submerged in a local shallow pond through the winterof 2009-2010. The bags of over-wintered seed were removed from the pond on April 5, 2010 and stored in arefrigerator until used on April 6. Subsequent sprouting tests conducted in small outdoor mesocosmsindicated that roughly 75% of these over-wintered seeds were viable.

The study plots were allowed to settle for 1 day after installation of the fenced exclosures to allow anydisturbed areas of the soft, flocculent sediment to reconsolidate and level out. At each of the three study sites,1 fenced plot and 1 open plot were randomly selected for seeding, leaving 1 fenced and 1 open plot at eachsite as non-seeded controls (Figure 6). On April 6, 2010, we hand-broadcast 16 oz. of drained, moist, over-wintered seed at each of the selected study plots. Water temperature at the time of seeding was 55° F.Broadcasted wet seed sank immediately upon landing in the water, thus minimizing any seed drift out of theseeded plots. We were careful to broadcast seed evenly over each plot area, including a halo that extendedabout 1 to 2 feet beyond the delineated plot boundaries. This resulted in a seeding rate of roughly 200 to 300lbs per acre in seeded plots. This seeding rate was 5 to 6 times higher than typically recommended forrestoration of wild rice stands (40 to 50 lbs per acre; Thompson and Luthin 2004). We chose to use this highseeding rate to ensure that there was a sufficient abundance of viable seeds in each seeded plot to producelush rice growth under favorable conditions.

7


Site Plot ID Plot Type Seeded Water Depth (in) Sediment

A A-F1 fenced 40 soft organic A-F2 fenced X 35 soft organic A-O1 open 40 soft organic A-O2 open X 40 soft organic

B B-F1 fenced X 38 soft organic B-F2 fenced 37 soft organic B-O1 open 44 soft organic B-O2 open X 43 soft organic

C C-F1 fenced X 27 soft organic C-F2 fenced 25 soft organic C-O1 open X 26 soft organic C-O2 open 25 soft organic

Table 1. Study plot identifiers, treatments, and physical characteristics. Reported water depth is mean plot depth(measured at all four plot corners) relative to the mean April–July water level (950.2 ± 0.2 ft; mean ± 1SD).

Figure 6. (a) Location of study sites in Upper Clam Lake and relative position of plots at each site. (b) Fenced plot(12"12 ft carp exclosure) with vertical fence extension; (c) open plot with no fence.

Site ASite C

Site B

O1

F2*O2*

F1

O1

F2

O2*F1*

O1*F2

O2

F1*

N!

0.5 mi

Open Plot (no fence)Fenced Plot

* Seeded

a

c

b

8


Spring Stem Counts: Floating-leaf StageWe assessed the abundance of wild rice (stem density, stems/m2) in each study plot on May 27, 2010. Thesespring surveys were conducted to document the survival and growth of wild rice through the floating-leafstage. We conducted non-destructive stem counts from a boat positioned next to each plot. Wild rice stemabundance in each of the 12 study plots was assessed using a 3-sided, square, 0.1 m2 PVC quadrat (32 " 32 cm)and an underwater video camera (Aqua-Vu Z-series; Outdoors Insight Inc., Crosslake, MN) mounted on anextendable 12-ft pole (Figure 8a). The quadrat frame was tethered to the end of the camera pole with a 5-ftpiece of high-visibility rope. This setup allowed for rapid placement and retrieval of the quadrat and alsominimized bias in its placement, as we could not see the quadrat or sediment underwater while it was beingplaced. Furthermore, this method allowed us to assess the plots without having to wade or swim in each plot,thus minimizing disturbance of the sediments and plants within each plot.

We used a stratified-random sampling approach to assess each plot, first dividing the plot into 9 equal-sizedcells (3 " 3 grid; each cell 4 " 4 ft), followed by random quadrat placement in 8 of these cells. Quadratplacement was randomized within each cell by suspending the quadrat over the center of each cell, gentlylowering the quadrat to the sediment with the open end pointing upward, and allowing the quadrat to tip toeither side. After the quadrat was in place, the underwater video camera was submerged until the quadratframe was clearly visible on the viewing screen (Figure 8b). In the field, live video was viewed on a 9-in LCDscreen while the camera was moved around the perimeter of the quadrat. All wild rice stems that originatedfrom the sediment area within the quadrat boundaries were counted while in the field, however accurate fieldcounts were somewhat difficult to perform due to poor screen-viewing conditions. Accordingly, we alsorecorded the underwater video from each quadrat assessment and noted the video time stamp at thebeginning and end of each assessment on the field data sheet. We later reviewed the recorded underwatervideo under better viewing conditions on a larger monitor to verify and update our field stem counts. Wedivided the final stem counts by the quadrat area (0.1 m2) to yield stem density (stems/m2) for each quadratsample.

To determine whether wild rice stem density in plots was affected by the presence or absence of carp, and bymanual seeding, we used a randomized complete block design analysis of variance (ANOVA) on mean stemdensity, with carp (present or absent) and seed (seeded or not seeded) as fixed effects, with site location (A, B,or C) as a random (block) factor (Zar 2010). If effects were significant (P < 0.05), we compared mean stemdensity between specific plot treatments using post hoc Student-Newman-Keuls (SNK) tests (#=0.05). Allstatistical analyses were performed using R statistical software, version 2.12.0 (R Development Core Team2010), along with the nlme R-package for ANOVA models (Pinheiro et al. 2010), and the GAD R-package forpost hoc SNK analyses (Sandrini-Neto and Camargo 2010).

a b Figure 8. (a) Equipmentused to perform stem countsand visual inspections of theenclosures; (b) image fromunderwater video recordedduring a quadrat stem count.

9


Mid-Summer Stem Counts: Emergent StageWe conducted a second assessment of each study plot on July 21, 2010. These mid-summer surveys wereintended to document the survival and growth of wild rice through the emergent stage. For each plot, weonce again performed non-destructive stem counts from a boat, using the same stratified-random approachdescribed earlier, with the slight modification of including counts from all 9 cells in each plot. The abundanceof emergent wild rice stems in each of the 12 study plots was assessed using the same methodology asdescribed for the spring surveys, except we used a floating 3-sided, square quadrat (0.1 m2; 32 " 32 cm) andvisually counted stems that emerged from the surface of the water within the placed quadrat frame (cameranot used). We then divided these surface stem counts by the quadrat area (0.1 m2) to yield stem density(stems/m2) for each quadrat sample. Statistical analysis of mid-summer stem density was identical to theanalyses used for the spring surveys.

Late-Summer Seed-head CountsWe conducted a third assessment of each study plot on August 31, 2010. These late-summer assessmentswere intended to (1) document the number and average mass of wild rice seed-heads produced within eachplot, and (2) determine whether mature seeds were produced on the seed heads. For each plot, we removed,enumerated, and weighed (fresh) all rice seed heads (whole plot), and then calculated the average mass perseed head for each plot. In addition, we noted whether mature seeds were found on seed heads from eachplot.

ResultsSpring Stem Density: Floating-leaf StageANOVA results indicated that in May, wild rice stem density in the study plots was significantly affected byboth carp (P < 0.0001) and seeding (P < 0.0001), with significant interaction between carp and seeding (P <0.0001); Table 2). Post hoc SNK tests indicated that mean stem density in the fence+seed plots (300 ± 21stems/m2; mean ±SE) was much greater than in all other plots (P < 0.01). By contrast, post hoc tests comparingother plot treatments indicated that there was no difference in mean stem density between the open plots(seeded/non-seeded; P > 0.05), or between non-seeded plots (fenced/open; P > 0.05). Open plots andunseeded plots had very low mean stem density (<1 stem/m2) in May, with 7 out of 9 plots having no wild ricestems at all (Figure 9a). Nearly all of the rice plants observed in plots during this time were in the floating-leafgrowth-form (Figure 11), with only a few individual plants showing early signs of the emergent growth-form.We also observed heavy carp activity (jumping and splashing) in shallow areas around each of the study sitesduring the survey.

Mid-Summer Stem Density: Emergent StageResults from the July surveys were nearly identical to those reported for the spring (May) surveys. ANOVAresults indicated that in July, stem density in study plots was significantly affected by both carp (P < 0.0001)and seeding (P < 0.0001), with significant interaction between carp and seeding (P < 0.0001; Table 2).Although mean wild rice stem density in the fence+seed plots during the July surveys (120 ± 12 stems/m2)was substantially lower than seen in May (300 ± 21 stems/m2), these plots all exhibited lush growth,suggesting that the observed reduction in stem density was due to self-thinning. Post hoc SNK tests indicatedthat July mean stem density in fence+seed plots was greater than mean in all other plot treatments (P < 0.01).By contrast, post hoc tests comparing other plot treatments indicated that in July, there was no difference in

10


mean stem density between the open plots (seeded/non-seeded; P > 0.05), or between non-seeded plots(fenced/open; P > 0.05). These plots all had very low average stem densities (<1 stem/m2) in July, with 8 out of9 plots having no wild rice stems at all (Figure 9a). All of the rice plants observed in plots during the Julysurveys were in the emergent growth-form (Figure 12), with many plants having flowering seed heads. Wealso observed a moderate amount carp activity (signs of carp movement) in shallow areas immediatelyadjacent to shore during the July survey.

Table 2. Randomized complete block design ANOVA results for May and July wild rice stem density (stems/m2) in study plots.Comparison of carp (present or absent) and seed (seeded or not seeded) as fixed effects, with site location (A, B, or C) as a random(block) factor.

Response variable DF F-value P-value

May Carp (present or absent) 1 137.4 <0.0001 Seed (seeded or not seeded) 1 138.2 <0.0001 Carp " Seed 1 138.2 <0.0001

July Carp (present or absent) 1 81.6 <0.0001 Seed (seeded or not seed) 1 79.7 <0.0001 Carp " Seed 1 79.7 <0.0001

Figure 9. Mean wild rice stem density (stems/m2) in Upper Clam Lake study plots observed on (a) May 27, 2010, during the floatingleaf stage of growth, and (b) July 21, 2010, during the emergent stage of growth. Vertical bars represent +1SE. Plot treatments: “Fence”denotes plots where carp were excluded, “Open” denotes plots where carp were allowed, and “+ Seed” denotes plots that weremanually seeded. Stem density in the three Fence+Seed plots was greater than in all other plots in both May and July (P < 0.01; posthoc SNK test).

Wild Rice Stem Counts in Study Plots: July 2010

0

50

100

150

200

250

300

350

400

Fence + Seed Fence Open + Seed Open

Wild

Ric

e S

tem

s / m

2

Site ASite BSite C

Wild Rice Stem Counts in Study Plots: May 2010

0

50

100

150

200

250

300

350

400


Wild

Ric

e S

tem

s / m

2

Site ASite BSite C

May July

(a) (b)

11


Late-Summer Seed-head CountsWith the exception of one wild rice seed-head found in a fenced, non-seeded plot, all other seed-heads werefound in the fence+seed plots (Figure 12). Although we observed heavy production of seed-heads at all threefence+seed plots during the July stem-count survey (flowering seed-heads), a severe storm in Augustdestroyed (uprooted) the rice at site A before we harvested seed heads. Consequently, we were only able tocollect seed-heads from the remaining two sites. The fence+seed plot at site B contained 1250 seed-heads(!90 seed-heads/m2) with a mean mass of 1.1 g per seed-head. The fence+seed plot at site C contained 750seed-heads (!60 seed-heads/m2) with a mean mass of 1.4 g per seed-head.

We found some mature seed being produced at both sites B and C, however, most seed-heads only had a fewmature seeds attached. Most seed heads at site B had mostly empty hulls and immature (milky) seedsattached, while seed-heads at site C had somewhat more mature seed. We were not able to determine howmuch mature seed had already dropped from seed heads prior to harvesting. Seed predation by worms mayalso have contributed to the low abundance of mature seeds. We observed a substantial number of riceworms feeding on seed heads at site B, with very few worms found at site C.

DiscussionOur study clearly showed that direct carp effects severely limited wild rice survival, growth, and reproductionin Upper Clam Lake. Furthermore, the high survival and growth of rice in the fence+seed plots, which weresubjected to the same water conditions as the other plots, suggested that indirect carp effects (increasedturbidity and nutrients) did not substantially impact wild rice in our study plots. Previous studies have shownthat abundant carp can severely reduce submersed aquatic vegetation in shallow areas through directuprooting or herbivory (Cahn 1929; Crivelli 1983; Sidorkewicj et al. 1998; Evelsizer and Turner 2006; Miller andCrowl 2006; Bajer et al. 2009). Our findings strongly suggest that these direct carp effects can also impact wildrice, despite being an emergent plant.

Although wild rice survival and growth can also be reduced by high water levels in June and July (Moyle 1944),our results clearly show that water level did not substantially affect wild rice survival or growth at our studysites, as evidenced by the lush rice growth in fence+seed plots. Water elevation in Upper Clam Lake fluctuatedsomewhat during the period from April through July 2010 (950.2 ± 0.2(SD) ft; Figure 10), but we did not seeany evidence of “drowning” rice, as described by Moyle (1944). Water elevation did increase dramatically inmid-August (+1.2 ft) after a large storm. The combination of high water and strong wind during and in theweek following this storm uprooted emergent rice in the fence+seed plot at site A, however, rice in plots atthe other two sites was not noticeably affected.

We found that direct carp effects severely reduced wild rice during its early stages of growth, as indicated bythe near complete lack of young wild rice shoots in open plots during the May survey. However, we were notable to determine whether carp had eaten the seeds, eaten young shoots, or merely uprooted new shootsbefore the end of May. Several papers report that carp actively eat plant seeds and shoots of some submersedaquatic plants (Crivelli 1983; Evelsizer and Turner 2006; Miller and Provenza 2007), however, the palatability ofwild rice seeds and shoots to carp has not been evaluated. Miller and Provenza (2007) reported that otheremergent aquatic plants, Typha (cattail) and Scirpus (bulrush), were unpalatable to carp due to the toughness(structural deterrent) of the plants, despite being high in nutritional value. This suggests that the similarlytough emergent form of wild rice would also be unpalatable to carp. However, young wild rice shoots aremuch less tough than mature plants, possibly making these shoots particularly vulnerable to carp herbivoryand uprooting during the spring period of heavy carp activity and spawning in shallow areas. Because of thenear complete absence of rice in our open plots in May and July, we were not able to evaluate whether carp

12


negatively affected rice plants once they reached the emergent stage. Regardless of whether carp eat wild riceseeds and shoots or merely uproot young plants, carp damage to rice during its early-season growth periodlikely represents a bottleneck that limits the amount of rice that reaches the emergent growth stage.

In addition to the rice growth within the study plots, we also found a few individual rice plants growing justoutside of the seeded plots (fenced and open) at site C in both May and July (Figures 11 and 12). The survivalof these plants may be attributable to sediment disturbance that occurred while installing the exclosures andopen site markers. Only the plots at site C were installed by wading; plots at sites A and B were installedentirely from a boat. At the time of plot installation, the downstream dam had not yet been closed, so thewater level was still at the drawn-down winter lake level (roughly 1 ft below the spring–summer waterelevation). Consequently, water depth at site C was too shallow to use a boat during study plot installation.While wading, we were careful not to disturb the sediment within each plot area, but we sank 2 to 3 feet intothe soft sediment while walking along the perimeter of each plot during the installations. Although weallowed the plots to settle for a day before seeding, it is possible that we left deep footprints that persistedaround each plot. Seeds that fell into persistent depressions in the sediment would have been somewhatmore protected from carp activities, giving them relatively more time to grow and establish roots before beingwithin reach of the carp. We can not be sure that this was indeed the reason for the observed wild rice growthoutside of the plots, however, it would be interesting to evaluate whether wild rice survival in carp-infestedlakes could be enhanced by creating footprint-like holes in the sediment just prior to seeding.

Figure 10. Water elevation in Upper Clam Lake during the 2010 wild rice plot study. Mean water elevation for April–July was 950.2 ±0.2 (SD) ft; August 950.6 ± 0.6 ft. Prior to April 15, the lake was refilling from the winter draw-down elevation (!949.0).

Upper Clam Lake Water Elevation: 2010

949.6

949.8

950.0

950.2

950.4

950.6

950.8

951.0

951.2

951.4

951.6

951.8

Apr-1 May-1 Jun-1 Jul-1 Aug-1 Sep-1

Wat

er E

leva

tion

(ft)

13


In addition to showing the strong impact of direct carp effects on wild rice, our study also clearly showed thatthe abundance of wild rice seeds remaining in sediments of Upper Clam Lake from past years was notsufficient to produce substantial rice growth in the absence of carp. The abundance of rice growth in non-seeded, fenced plots was very low throughout the study (0 stems/m2 in May; <1 stem/m2 in July), and was notsignificantly greater than we observed in any of the open plots. Moyle (1944) reported that although wild ricestands may experience periodic natural failures, there is typically sufficient seed production during thesefailures to allow the stand to recover in the subsequent year. Similarly, Walker et al. (2010) reported that wildrice stands exhibit natural cycles of failure followed by recovery, roughly every four-years. Based upon (1) the4-year duration of the current rice decline in Upper Clam Lake, (2) the low number of seeds found during the2009–2010 seed enumeration study, and (3) the lack of rice growth in our non-seeded fenced plots, the failureof rice in Upper Clam Lake does not appear to be a typical natural decline.

Collectively, these findings suggest that restoration of wild rice stands in Upper Clam Lake will require (1) adramatic reduction of carp abundance to less than !30 lbs of carp/acre (Bajer et al. 2009), and (2) manualseeding of near-shore areas. Furthermore, our findings suggest that neither of these strategies alone wouldresult in a rapid recovery of rice beds in Upper Clam Lake. If possible, any manual seeding of wild rice shoulduse seed collected from remaining sparse stands of rice in Upper Clam Lake, or from a nearby lake (e.g. LongLake). Recent genetic studies have shown that isolated stands of wild rice adapt to local conditions over time,resulting in unique genetic strains of rice that are particularly suited to thrive in given bodies of water (Kern A.,Associate Professor of Biology, Northland College – Ashland (WI); Sept 2010, pers. comm.). Reseeding UpperClam Lake with locally produced seed would help to preserve the locally adapted strain of rice, and may helpto improve the persistence of any restored stands in the lake.

14


Figure 11. Upper Clam Lake study plots; wild rice growth at the time of the spring stem-count surveys (May 27, 2010; floating-leaf stage).

Fence + Seed Fence (no seed) Open + Seed Open (no seed)

A

C

B

15


Figure 12. Upper Clam Lake study plots, showing wild rice growth at the time of the late-summer stem-count surveys (July 21, 2010; emergent stage).

Fence + Seed Fence (no seed) Open + Seed Open (no seed)

A

C

B

16


References

Bajer P, Sullivan G, Sorensen P. 2009. Effects of a rapidly increasing population of common carp on vegetativecover and waterfowl in a recently restored Midwestern shallow lake. Hydrobiologia. 632:235–245.

Cahn AR. 1929. The effect of carp on a small lake: the carp as a dominant. Ecology. 10:271–274.

Crivelli, AJ. 1983. The destruction of aquatic vegetation by carp. Hydrobiologia. 106:37–41.

Evelsizer VD, Turner AM. 2009. Species-specific responses of aquatic macrophytes to fish exclusion in a prairiemarsh: a manipulative experiment. Wetlands 26:430–437.

Johnson, JA. 2010. Wild Rice Seed Enumeration Report: 2009-2010, Upper Clam Lake, Lower Clam Lake , LongLake , and Clam River Flowage. Final report to SEH Inc – Spooner (WI). Freshwater Scientific Services LLC –Maple Grove, MN

King DR, Hunt GS. 1967. Effect of carp on vegetation in Lake Erie Marsh. J. Wildl. Manage. 31:181–188.

Miller SA, Crowl TA. 2006. Effects of common carp (Cyprinus carpio) on macrophytes and invertebratecommunities in a shallow lake. Freshw. Biol. 51:85–94.

Miller SA, Provenza FD. 2007. Mechanisms of resistance of freshwater macrophytes to herbivory by invasivejuvenile common carp. Freshw. Biol. 52:39–49.

Moyle JB. 1944. Wild rice in Minnesota. J. Wildl. Manage. 8:177–184.

Moyle JB, Kuehn JH. 1964. Carp, a sometimes villain. p. 635–642 In: Linduska JP and Nelson AL (eds.), WaterfowlTomorrow. U.S. Dept. Int., Bur. Sport Fish. And Wildl., Fish Wildl. Ser. Washington, DC.

Pinheiro J, Bates D, DebRoy S, Sarkar D, and the R Development Core Team. 2010. nlme: Linear and NonlinearMixed Effects Models. R package version 3.1-97.

R Development Core Team. 2010. R: A language and environment for statistical computing. R Foundation forStatistical Computing. Vienna (Austria). ISBN 3-900051-07-0. URL http://www.R-project.org/.

Sandrini-Neto L, Camargo MG, and the R Development Core Team. 2010. General ANOVA Design (GAD):Analysis of variance from general principles. R package version 1.0.

Sidorkewicj NS, Cazorla ACL, Murphy KJ, Sabbatini MR, Fernández OA, Domaniewski JCJ. 1998. Interaction ofcommon carp with aquatic weeds in Argentine drainage channels. J. Aquat. Plant Manage. 36:5–10.

Thompson AL, Luthin CS. 2004. Chapter 12: Wild Rice Community Restoration, p. 117–121 In: Griffin MPA andWatermolen DJ (eds.), Wetland Restoration Handbook for Wisconsin Landowners, 2nd Ed. WisconsinDepartment of Natural Resources, Bureau of Integrated Science Services. Madison (WI).http://dnr.wi.gov/wetlands/handbook.html

Walker RD, Pastor J, and Dewey BW. 2010. Litter quantity and nitrogen immobilization cause oscillations inproductivity of wild rice (Zizania palustris L.) in northern Minnesota. Can. J. of Bot. 84:485–498.

Zar JH. 2010. Biostatistical analysis, 5th ed. Upper Saddle River (NJ). Pearson Prentice Hall.

17


Appendix A – Observations for Other Plant Taxa

Although our study focused on evaluating the effects of carp on wild rice, we also recorded the occurrence ofother aquatic plant taxa (hereafter “plants”) in the study plots. Although we have included a summary of theseobservations in this appendix, more comprehensive evaluations of carp effects on submersed aquatic plantshave been conducted by others (Cahn 1929; Crivelli 1983; Sidorkewicj et al. 1998; Evelsizer and Turner 2006;Miller and Crowl 2006; Bajer et al. 2009).

In May, we generally observed low plant frequency (! 50%) and low plant abundance (very sparse growth) inthe study plots, with no apparent pattern to differences in frequency between plot treatments (Figure A-1 andFigure 12). However, the frequency of plants in May differed somewhat between sites. At site A, only one plot(non-seeded, fenced) had plants, but at sites B and C, all but one plot had plant frequency " 25%. In July, plantfrequency was generally much higher at site C than at the other two sites. Although plant frequency was 100%in 3 of 4 plots at site C in July, the abundance and diversity of plants appeared to be greatest in the non-seeded, fenced plot (Figure 12). We did not directly measure plant abundance (biomass or stem counts),however, plant growth was clearly more abundant and more diverse within the non-seeded, fenced plot atsite C than in the area immediately outside of the plot (Figure A-2). This suggests that the survival and growthof plants at this shallow site was enhanced by excluding carp. We did not see a similar enhancement of plantsin any of the fence+seed plots. The lush wild rice in these plots appeared to suppress the growth of otheraquatic plants, likely due to direct competition for light or nutrients (Moyle 1944).

Figure A-1. Frequency (% occurrence) of submersed aquatic plant taxa (combined occurrence of all taxa other than wild rice) in UpperClam Lake study plots (May and July). Plot treatments: “Fence” denotes plots where carp were excluded, “Open” denotes plots wherecarp were allowed, and “+ Seed” denotes plots that were manually seeded with wild rice.

Figure A-2. Non-seeded, fenced plot at site C (July 21, 2010). Note difference in density and diversity of plant growth inside andoutside of the fenced plot. See Figure 12 for additional comparisons with other plots and sites.

Other Plant Taxa in Study Plots: May 2010

0

10

20

30

40

50

60

70

80

90

100


% O

ccur

renc

e

Site ASite BSite C

Other Plant Taxa in Study Plots: July 2010

0

10

20

30

40

50

60

70

80

90

100


% O

ccur

renc

e

Site ASite BSite C

May July

18


Table A-1. List of additional plant taxa found in each plot and site in 2010. Plant frequency (% of quadrat samples in each plot)reported to nearest 10%. Abundance (biomass) not directly measured, but is described qualitatively (Low, Med, High).

Site Plot Taxon Frequency (%) AbundanceFence+Seed Nymphaea odorata 10 LowFence Ceratophyllum demersum

Potamogeton pusillusMyriophyllum sibiricum

Present *5050

LowLowLow

Open+Seed Nymphaea odorataVallisneria americana

1010

LowLow

A

Open None — —

Fence+Seed Ceratophyllum demersum 30 LowFence Ceratophyllum demersum

Potamogeton pusillusPotamogeton nodosus cf.

103010

LowLowLow

Open+Seed Nymphaea odorata Present Low

B

Open Ceratophyllum demersumMyriophyllum sibiricum

3010

LowLow

Fence+Seed Potamogeton pusillusCeratophyllum demersum

30Present

LowLow

Fence Ceratophyllum demersumElodea canadensis

Myriophyllum sibiricumNymphaea odorata

Potamogeton pusillusPotamogeton zosteriformis

Zosterella dubia

1010

Present201001030

LowLowLowMedHighLowMed

Open+Seed Ceratophyllum demersumNajas flexilis

Potamogeton pusillusVallisneria americana

Zosterella dubia

40901001010

MedMedMedLowLow

C

Open Ceratophyllum demersumMyriophyllum sibiricum

Najas flexilisPotamogeton crispusPotamogeton pusillus

Zosterella dubia

101080203010

MedLowMedLowLowLow

* Present denotes taxa that were observed in the plot, but were not encountered in any quadrat samples.

Documents

Effects of Carp on the Survival and Growth of Wild Rice in ... · Final report to St. Croix Tribal Environmental Services – Natural Resources Department, Webster (WI). Freshwater