- Email: [email protected]
Baldcypress and water tupelo responses to insect defoliation and nutrient augmentation in Maurepas Swamp, Louisiana, USA
Forest Ecology and Management 236 (2006) 295–304 www.elsevier.com/locate/foreco
Baldcypress and water tupelo responses to insect defoliation and nutrient augmentation in Maurepas Swamp, Louisiana, USA Rebecca S. Effler *, Richard A. Goyer, Gerald J. Lenhard LSU AgCenter, Department of Entomology, Louisiana State University, 404 Life Sciences, Baton Rouge, LA 70803, USA Received 28 February 2006; received in revised form 12 September 2006; accepted 13 September 2006
Abstract Forested wetlands in Louisiana are hydrologically isolated from the Mississippi River, impacted by saltwater intrusion, and are sinking, resulting in more frequent flooding for longer periods. These problems are exacerbated in some areas by impoundment from man-made structures. Additionally, defoliation of the two dominant trees, baldcypress (Taxodium distichum L. Rich) and water tupelo (Nyssa aquatica L.), frequently occur during spring. In Louisiana, the baldcypress leafroller, BCLR (Archips goyerana Kruse) and the forest tent caterpillar, FTC (Malacosoma disstria Hubner) annually defoliate an average of 120,000 ha of baldcypress–tupelo swamps. In combination, these factors have led to severely degraded forested wetlands with reduced basal area, dieback and tree mortality over extensive, formerly productive swamplands. Restoration plans call for reintroducing Mississippi River water by diversions, to increase sediment elevation, promote natural regeneration, and enhance primary and secondary productivity. Nutrient augmentation of baldcypress and water tupelo trees in Maurepas Swamps (similar to NO3 loading rates expected from 229 m3/s diversion) increased radial growth (cm2) of both species (especially baldcypress) in degraded forest stands. Nutrient augmentation also increased nitrogen in foliage (insect frass (feces), spring clipped leaves, and abscised leaf litter) for both tree species. Nitrogen content of canopy foliage and leaf litter was positively correlated with site health. These findings support hypotheses that swamps in southeastern Louisiana are nutrient limited, and existing trees can utilize, benefit, and act as nutrient sinks for nutrient-laden river water accompanying diversions. Defoliation of these swamp species, although not uniform among years, did not result in losses in radial growth, but increased overall litter deposition for baldcypress. An increase in nitrogen content in insect frass in response to nutrient augmentation indicates defoliator populations might increase as a result of nutrient influxes. # 2006 Elsevier B.V. All rights reserved. Keywords: Diversions; Taxodium distichum; Nyssa aquatica; Malacosoma disstria; Archips goyerana; Lepidoptera; Nitrogen; Litterfall
1. Introduction Forested wetlands are a valuable natural resource that stimulates and sustains both economic productivity and ecological functioning on local and global scales. They have been identified as important nutrient sinks that enhance water quality and export organic debris to many aquatic food webs (Mitsch and Gosselink, 1993). However forested wetlands are being lost worldwide due mainly to altered natural hydrology. In Louisiana, USA, freshwater input from the largest North American river system (Mississippi River) has been reduced in or in some cases completely isolated from forested wetlands by
* Corresponding author. Present address: University of Georgia Marine Institute, Sapelo Island, GA 31327, USA. Tel.: +1 912 485 2225; fax: +1 912 485 2133. E-mail address: [email protected] (R.S. Effler). 0378-1127/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.foreco.2006.09.014
levees and dams (Louisiana Coastal Wetlands Conservation, 1998). Wetlands that were historically influenced, but now isolated from these riverine inputs, are currently subsiding at approximately 1 m/century (Salinas et al., 1986), and decreasing in productivity and biodiversity (Conner and Day, 1988; Conner and Buford, 1998). Many swamps in coastal Louisiana, including Maurepas Basin, are hydrologically isolated from the Mississippi River (Mossa, 1996), predominantly impounded by man-made structures (Turner, 1997), degraded by saltwater intrusion (Conner and Brody, 1989; Conner and Toliver, 1990), and have experienced more frequent flooding, for longer periods (Conner and Day, 1984). To sustain current canopy trees and promote natural regeneration, Mississippi River inputs (nutrient flows and other riverine inputs by diversions) need to be re-established into declining wetland systems (Baumann et al., 1984; Templet and Meyer-Arendt, 1988; Boesch et al., 1994; Day et al., 2000, 2003). Active and
296
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304
approved diversions are aimed to flush out soil phytotoxins and interstitial salt accumulation from saltwater events, and increase elevation and production as historically accomplished by natural hydrological conditions (Louisiana Coastal Wetlands Conservation, 1993, 1998). Two diversions in operation (Atchafalaya Spillway and Caernarvon) have demonstrated that riverine inputs into marshes result in increased primary and secondary production (Visser, 1989; Louisiana Coastal Wetlands Conservation, 1998; Lane et al., 1999). In addition to abiotic stressors present (impoundment, subsidence, increased salinity, nutrient deprivation) in southeast Louisiana swamps, the two dominant tree species, water tupelo (Nyssa aquatica L.) and baldcypress (Taxodium distichum L. Richard), are subject to frequent springtime defoliation. This defoliation event simultaneously occurs for the two species and starts in March and continues through mid May. Water tupelo is defoliated by the forest tent caterpillar, FTC (Malacosoma disstria Hubner). Defoliation levels in south Louisiana alone have exceeded 250,000 ha during a single season (Nachod and Kucera, 1971; Nachod, 1977) and average 120,000 ha each year with much occurring in Maurepas Basin (Goyer and Chambers, 1996). Abrahamson and Harper (1973) found that up to 45% of the annual incremental growth of water tupelo was lost due to severe FTC defoliation during a 5-year impact study in Alabama. In 1983, an epidemic of the baldcypress leafroller, BCLR (Archips goyerana Kruse) occurred on baldcypress in and around the Atchafalaya Basin (Goyer and Lenhard, 1988). This outbreak has since spread to southern Lake Maurepas and Lake Verret (Goyer and Chambers, 1996) defoliating as much as 200,000 ha in a single season and averaging 72,000 ha each year. Baldcypress, renowned for its lack of serious insect and disease problems (Brown and Montz, 1986), presently experiences defoliation damage resulting in significant reduction in radial growth and dieback (Braun et al., 1989; Allen et al., 1997; Goyer and Chambers, 1996). Baldcypress health generates concern among biologists because it is one of the most ecologically valuable tree species in the southeastern floodplain forests (Brown and Montz, 1986). Baldcypress maintains forest ecosystem integrity through its extensive root system, canopy, and long-lived nature. It is the keystone habitat species for many game animals, sport fish, and formerly or presently threatened and endangered species such as the bald eagle (Haliaeetus leucophalus (L.)), Louisiana black bear (Ursus americanus luteolus Griffith), American alligator (Alligator mississippiensis (Daudin)), Florida cougar (Felis concolor coryi (Bangs)) (Dennis, 1988), and potentially the recently rediscovered ivory-billed woodpecker. Forest defoliation by insects can also lead to severe disruptions in nutrient cycling by altering biomass, detrital processes, and nutrient content of litterfall reaching the forest floor. Studies have shown that defoliation products (foliage, insect tissues, and frass) transfer a considerable amount of nutrients to the litter layer (Schowalter, 1981; Schowalter and Crossley, 1983; Seastedt and Crossley, 1984; Swank et al., 1981). Premature litterfall resulting from artificial defoliation returned 10 times as much nitrogen to the litter as would normal
litterfall (Klock and Wickman, 1978). Schowalter et al. (1991) found that defoliators removing 20% of the foliage from young Douglas-fir (Pseudotsuga menseii (Miirb.) Franco), doubled the amount of precipitation reaching the forest floor, and increased N, K, and Ca flow to the forest floor by 20–30%. With more nutrients reaching the forest floor, defoliation may affect water quality in adjacent streams (Coulson and Witter, 1984; Klock and Wickman, 1978; Swank et al., 1981; White and Schneeberger, 1981). In the northeastern United States, herbivory does not appear to significantly affect water quality (Bormann and Likens, 1979; White and Schneeberger, 1981), perhaps because of the high baseline concentration of nutrients in these streams (Swank et al., 1981). However, nitrateimpoverished streams in the southeastern United States show significant increases in NO3-N concentrations during periods of defoliation (Swank et al., 1981). Hutchens and Benfield (2000) found that breakdown rates of chestnut oak (Quercus prinus L.) leaves were faster for secondflush leaves produced after gypsy moth (Lymantria dispar L.) defoliation than those of natural spring-flush leaves shed in autumn in three of six streams tested. They concluded that insect defoliation accelerates detritus processing in southern Appalachian streams. However, induced defenses by the plant as a result of insect pressure may reduce litter decomposition rates (Zlotin and Khodashova, 1980). Based on N-15 labeled isotopes, N-15 in gypsy moth frass was mobilized more quickly (40% incorporated in the soil) than N-15 in leaf litter (80% remained undecomposed in the leaf litter) in an oak forest (Christenson et al., 2002). However, frass N may be largely unavailable to plants and microorganisms, as little N-15 was found in extractable, microbial, or readily mineralizable pools. These processes may be particularly important for declining, nutrient-limited Louisiana swamps, as litterfall associated with insect herbivory may be an important nutrient source for primary and secondary productivity in the spring when biological emergence is at its peak. Importantly, baldcypress–tupelo swamps are known to be long-term sinks for both nitrogen and phosphorous via burial of partially decomposed organic and inorganic matter under reduced soil conditions and denitrification (Kemp et al., 1985). However, little research has been conducted into nutrient/biological interactions of the trees and defoliators in these swamps, with the exception of Meeker (1992) and Johnson (2004) that address quality of baldcypress foliage and insect performance. During BCLR larval activity, nutrient concentration (N, P, K, and Mg) of baldcypress foliage decreased as feeding progressed; and constituents representing tree defense mechanisms increased (Meeker, 1992), rapidly creating foliage characteristic unfavorable for the BCLR. Meeker (1992) also found that a flooded swamp may have more suitable foliage for the larvae, as suggested by having higher total nonstructural carbohydrates in foliage than in an unflooded swamp. Objectives of this study were to determine if nutrient enhancement of mature baldcypress and water tupelo trees will likely (1) increase stem growth, litter biomass, and foliar nitrogen content of trees in swamps exhibiting various states of decline, and (2) compensate for basal area growth loss as a result of defoliation by insects. Also, (3) determine the
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304
influence of defoliation on annual biomass of litterfall of mature baldcypress and water tupelo trees in these areas. 2. Methods 2.1. Description of Maurepas Swamps Maurepas Swamps are located in the Lake Pontchartrain Basin in southeast Louisiana (Fig. 1). Lake Maurepas (308100 N and 918300 W) is a large lake roughly 21 km in diameter. At the turn of the 20th Century, swamps in this basin, as well as most of coastal Louisiana, were harvested en masse (Mancil, 1980). Maurepas forested wetlands are the second largest contiguous coastal forest and consist of 77,550 ha of swamps and 5204 ha of fresh/intermediate marshes (LACOAST 2050, 1998). Swamps along the southern shore of Lake Maurepas are composed of 80% baldcypress and 20% water tupelo trees (Wilson et al., 2001), while deeper areas west of the lake show the reverse composition. The key driving factors of forest decline are saltwater intrusion, land subsidence, and lack of nutrient inputs (Thompson, 2000; Wilson et al., 2001; Lane et al., 2003; Shaffer et al., 2003). Insect defoliation in addition to these stressor further declines baldcypress and water tupelo health. The BCLR has recently reached outbreak proportions in southern Maurepas Swamps (1996–1999). Populations of FTC have been locally epidemic throughout the swamps since 1948, when records were kept. This swamp system is a candidate
297
hydrological restoration area (via river diversions) because of its rapid loss of swamp habitat and its previous history of receiving inputs from the Mississippi River. Furthermore, most nutrients accompanying the proposed diversion for the area are expected to be retained in the swamp and should not adversely affect water quality in Lake Maurepas (Lane et al., 2003). 2.2. Experimental design The Lake Maurepas field study to evaluate effects of nutrient augmentation on tree above-ground productivity was conducted from December 2000 to December 2003. The experiment consisted of a split-plot design. The main plot included three forest decline classes were (slight, moderate and severe). Locations of these decline classes are located in Fig. 1. Within each decline class, two plots were installed. Each plot consisted of a 2 3 factorial design. Within each decline class, two tree species (baldcypress and water tupelo) were assessed. Each tree species randomly received three nutrient regimes [no nutrients, a 1 dose of 236 g/m2 of Osmocote1 18-6-12 (B.W.I. Jackson, MS, USA), and a 2 dose]. The 1 rate is equivalent to a N loading rate of 229 m3/s springtime effluent of a Mississippi River diversion (Lane et al., 1999; Day et al., 2006). Each decline class had seven replicates (trees) of each fertilizer-species combination totaling 42 trees per decline class. The seven replicates were divided between the two plots in each decline class resulting in either three or
Fig. 1. Location of Maurepas Swamp and study sites.
298
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304
four replicate trees per plot. Trees were estimated to be between 40 and 80 years old. Trees were selected based upon minimum overlapping canopies to allow for litterfall collection. Osmocote fertilizer was applied each spring from 2001 to 2003 prior to bud break. Area of coverage was determined by estimating average canopy radius (dripline) to the nearest 0.5 m. The fertilizer was uniformly distributed equidistant around the base of the tree once a circular area was established. All areas selected also had recent defoliator activity on baldcypress and water tupelo. 2.3. Data collection Aluminum growth bands (1.5 cm 0.1 cm) were placed on all trees in January 2001 at breast height or at 0.5 m above butt swell on larger trees. Stem growth was monitored from January 2001 to December 2003 and recorded each year during dormancy (January), after defoliation (May), and late-summer (September) to the nearest 0.025 cm. A single litterfall/frass collection trap was placed under each tree. Litterfall traps were positioned 1 m above the forest floor and consisted of a wooden frame constructed from #2 treated lumber with a surface area of 0.25 m2 and 10 cm in depth. Trap bases were lined with 2 mm vinyl screen. Litterfall was collected once every 2 months, dried, and separated by plant species, and leaf material was weighed to the nearest 0.001 g. After refoliation occurred in June, leaves from the crown on the trees were sampled from approximately 10 to 15 m height by pole pruner or shotgun (clipped leaves). Subsamples of both clipped (spring) and leaf litter (natural abscission) were analyzed for N content. During spring (peak feeding activity of both insect species), these traps served as frass (feces) collection traps by inserting fine, saran screening (0.25 mm mesh) stretched over a 1.9 cm (3/4 in.) PVC pipe frame constructed to fit tightly (approximately the same area) inside the litterfall trap. Frass was collected once a week by sweeping the frass trap into a 120 ml vinyl specimen cup. Samples were returned to lab, uncovered, oven dried at 65 8C for 48 h, and then stored at room temperature until they could be separated from other litter and weighed to the nearest 0.001 g. During the last week of defoliation, percent defoliation of each tree was estimated to the nearest 10%. To more accurately measure annual leaf productivity across all defoliation levels, ‘‘frass’’ biomass was added to end of the year ‘‘leaf litter’’ (=‘‘litterfall’’) of each species and trends between the two were compared. In 2002 and 2003, subsamples of both clipped leaves and leaf litter were analyzed for N–P–K content (see SoutherEffler, 2004). For clarity, and its importance to both insect and tree nutrition, only N content is reported here. Nitrogen analyses were conducted on subsamples of frass (from peak collection period for each species) to evaluate nutrient inputs from insect herbivory independently from undefoliated, senesced leaves, which presumably contained a lower nutrient composition than frass. The 2002 frass was the only year that had representation of all treatment combinations. Thus, a second analysis was conducted in order to directly compare nutrient content of frass to litter and foliar parameters.
Nitrogen content of all foliage and frass was ascertained using AOAC (1995) protocols at the LSU/LDAF Agricultural Chemistry Laboratory on the LSU-Baton Rouge campus. 3. Statistical analyses Initially, all variables were analyzed as a split-plot design, as described in Section 2. However, plot variation in each decline class did not statistically differ (non-significant F-value < 1.7), so the degrees of freedom and sum of squares were pooled. Therefore, unless specifically stated, all variables were analyzed as a completely randomized design (CRD) with a 2 3 3 factorial arrangement: two species (baldcypress and water tupelo), three fertilizer treatments (0, 1, and 2), and three decline classes (slight, moderate, and severe decline). Fisher’s LSD was used for specific comparisons of interaction components. 3.1. Basal area growth An ANCOVA was conducted using the proc MIXED statement on SAS 8.0 (SAS Inst. Inc., 1999). The response variable was basal area growth (cm2/2 years) from January 2002 to December 2003. These data were analyzed as a CRD with a 2 3 3 factorial arrangement. To account for larger trees yielding higher basal area growth than smaller trees, initial stem basal area was added as a covariable. Two outliers were removed from this analysis. Further, analyses were conducted to test if fertilization compensated for loss of basal area growth resulting from defoliation. Yearly basal area growth from 2002 and 2003 of each species was regressed with defoliation from that year in each of the fertilizer treatments. Again, stem basal area was added as a covariable in each model. Each of the six models had 42 observations. 3.2. Litter productivity Two ANCOVAs were conducted on litter productivity, one on the biomass of leaf litter (fallen leaves only) and the other on the biomass of litterfall (leaf litter plus frass) (weighed to the nearest 0.001 g/0.25 m2/year). Defoliation (%) of each tree for that year served as a covariable in each model to test for the effects on leaf productivity. Initially, these variables were analyzed as a randomized block design (RBD) (with a block on the year— 2002 and 2003 collections) with a 2 3 3 factorial arrangement. However, variation due to year did not statistically differ for both analyses, so the degrees of freedom and sum of squares were pooled into the overall model. Therefore, the data were analyzed as a CRD with a 2 3 3 factorial arrangement. In addition, simple linear regression equations of the covariable (defoliation) were conducted for both leaf litter and litterfall. 3.3. Nitrogen content of clipped leaves and leaf litter 2002–2003 An ANCOVA was conducted on nitrogen (%) content with defoliation (%) as a covariable. Nitrogen (%) was analyzed as a
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304
randomized block design (RBD) (with a block on the year— 2002 and 2003 collections) with 2 2 3 3 factorial arrangement. The treatments included two leaf types (spring clipped leaves and leaf litter), two species (baldcypress and water tupelo), three fertilizer treatments (0, 1, and 2), and three decline classes (slight, moderate, and severe decline). Each leaf type was represented in each of the cross-classified species, fertility, and decline class categories. The number of trees sampled each year was n = 36 for litter and n = 122 for clipped for 2002, and n = 36 for leaf litter and n = 126 for clipped leaves for 2003. Degrees of freedom were Satterthwaite adjusted for unequal sample sizes. One outlier from the clipped leaves was removed. In addition, simple linear regression equations of the covariable (defoliation) were conducted for both leaf litter and litterfall. 3.4. Comparison of the nitrogen content of frass to other leaf parameters in 2002 As stated in Section 2, the 2002 data set was the only year frass was represented in all treatment combinations, and therefore a model was constructed for that year only, comparing relative amounts of nitrogen in frass to that in clipped leaves and leaf litter. An ANOVA was conducted on nitrogen (%) content as a CRD with a 2 3 3 3 factorial arrangement. The treatments were as above (Section 3.3) except that three leaf types (frass, spring clipped leaves, and leaf litter) were utilized. Each leaf type was represented in each of the cross-classified species, fertility, and decline class categories. Number of trees sampled for each leaf type was n = 101 for frass, n = 122 for clipped, and n = 36 for leaf litter. One outlier from the clipped leaves was removed. 4. Results Specific sites for this study represent three distinct degrees of decline in the Maurepas Basin—slight, moderate, and severe decline, based on a swamp productivity survey by Wilson et al. (2001). In this system, tree density is somewhat reflective of the varying degrees of forest decline. Basal area (BA) for each decline class in this study were slight decline class, BA = 32.84 1.14 m2/ha, moderate decline, BA = 15.72 0.94 m2/ha, severe decline BA = 16.73 0.90 m2/ha. Although basal area density of the severe decline class was similar to that of the moderate decline, the severe decline class had less canopy closure due to crown loss and dieback, especially water tupelo. These trees were missing the apical portion of the main stem resulting in mostly coppiced growth, indicative of the most degraded swamps in the basin where trees still exist. In addition, sites were also required to have recent insect defoliation of both tree species. Percent defoliation and diameter (cm) means and ranges of study trees are summarized in Tables 1 and 2. 4.1. Basal area growth Although the three-way interaction was not significant in the overall model, six a priori comparisons (using Fisher’s LSD) were used to determine if fertilization increased growth for each
299
Table 1 Diameter at breast height (cm) of study trees, by species for the decline class in 2002–2003 (n = 21 for each mean) Decline
Species
Mean (1 S.E.)
Minimum
Maximum
Slight
Water tupelo Baldcypress
24.9 1.0 31.9 1.9
16.5 14.5
34.5 48.3
Moderate
Water tupelo Baldcypress
30.3 1.1 25.2 1.7
21.6 11.7
41.9 46.0
Severe
Water tupelo Baldcypress
36.7 2.2 33.0 2.4
14.5 13.5
56.6 56.4
species in each decline class (Fig. 2). Results show that neither species benefited from fertilization application at the slight decline class; both species grew more with the fertilizer applications in the moderate decline class; and only baldcypress grew more with the fertilizer applications at the severe decline class while water tupelo did not. Basal area growth was highest in the moderate decline class (decline class, F 2,105 = 19.78, p < 0.0001, Fig. 2), which was mainly due to the differential utilization of fertilizer in each decline class (decline class fertilizer, F 4,105 = 3.51, p = 0.0099, Fig. 2). Fertilization in the moderate decline class led to increased tree growth compared to either the slight or severe decline class. Also, baldcypress trees grew more in basal area than water tupelo trees (F 1,105 = 23.99, p < 0.0001, Fig. 2). Defoliation (%) could not be correlated to basal area growth. 4.2. Litter productivity Fertilization did not significantly increase biomass of leaf litter (F 2,233 = 1.38; p = 0.2533). Leaf litter was highest under trees at the slight decline class, followed by the moderate decline class, and lowest at the severe decline class (Fig. 3, gray bars). This trend differed with regard to species; baldcypress trees produced significantly more leaves per tree at the slight Table 2 Percent defoliation of study trees, by species, for the decline class in 2002 and 2003 (n = 21 for each mean) Year/decline
Species
Mean (1 S.E.)
Minimum
Maximum
Water tupelo Baldcypress
99.5 0.5 77.9 4.8
90 30
100 100
Moderate
Water tupelo Baldcypress
12.4 4.6 72.9 6.1
1 15
70 100
Severe
Water tupelo Baldcypress
61.6 9.8 3.6 1.4
1 1
100 30
Water tupelo Baldcypress
99.3 0.5 62.6 6.4
90 10
100 100
Moderate
Water tupelo Baldcypress
2.7 0.9 71.7 6.3
1 20
20 100
Severe
Water tupelo Baldcypress
3.6 1.4 2.2 0.5
1 1
30 10
2002 Slight
2003 Slight
300
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304
Fig. 4. Baldcypress litterfall (frass and leaf litter) (g/0.25 m2/year) and percent defoliation of each tree for 2002 and 2003 (n = 126, 95% CI).
Fig. 2. Cumulative basal area growth (cm2/2 years) of (a) water tupelo and (b) baldcypress trees after receiving three levels of fertilization (1 = 236 g/m2 of Osmocote1 18-6-12) in three forest decline classes from 2002 to 2003 (n = 7 trees for each bar, 1 S.E.).
and severe decline classes, while water tupelo had a higher leaf litter than baldcypress at the moderate decline class (decline class species, F 2,232 = 5.07, p = 0.0070, Fig. 3, gray bars). As defoliation increased, a covariable in the model, biomass of leaf litter decreased (F 1,232 = 4.60, p = 0.0330). When defoliation was regressed by each species by decline class, only baldcypress at the moderate decline class was significant (r2 = 0.323, y = 0.304x + 57.986, F 1,40 = 19.07, p < 0.0001). Results for litterfall (leaf litter plus frass) showed that baldcypress produced significantly more litterfall than water tupelo at all decline classes (F 1,232 = 108.71, p < 0.0001, Fig. 3, white bars). Furthermore, baldcypress trees’ litterfall significantly decreased as forest health decreased; while water tupelo had similar litterfall productivity in the slight and moderate decline classes, but significantly less at the severe decline class (decline class species, F 2,232 = 5.07, p = 0.0070, Fig. 3). In this model, as defoliation increased, a covariable in the model, biomass of litterfall increased (F 1,232 = 4.60, p = 0.0330). Simple linear regressions of percent defoliation and litterfall showed a significant increase in baldcypress litterfall productivity as defoliation increased (Fig. 4), while tupelo did not. This is in contrast to the leaf litter analysis, where defoliation was negatively correlated to leaf litter productivity. 4.3. Nitrogen content of clipped leaves and leaf litter 2002–2003
Fig. 3. Mean annual litterfall (g/0.25 m2/year), leaf litter (leaf fall only – bottom gray) and litterfall (leaf litter + frass top white) of baldcypress (solid) and water tupelo (patterned) trees in three forest decline classes from 2002 to 2003 (n = 42 for each bar, 1 S.E., Fisher’s LSD, a = 0.05 for both leaf litter and litterfall).
As fertilization increased, foliar nitrogen levels significantly increased (main effect—F 2,28.9 = 7.76, p = 0.020, linear contrast—F 1,29 = 15.51, p = 0.0005, Fig. 5). Fertilizer application did not interact with leaf type, species, or decline class, indicating that foliar nitrogen increased as a result of fertilization in both spring clipped and leaf litter samples, in both baldcypress and water tupelo and in trees in each of the forest decline classes. As expected, spring clipped leaves had a higher nitrogen content than those in leaf litter (F 1,29 = 89.06,
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304
Fig. 5. Foliar nitrogen content of spring clipped leaves and leaf litter in response to fertilizer application to baldcypress and water tupelo trees (1 = 236 g/m2 of Osmocote1 18-6-12, a priori comparison of 0 vs. 2 using Fisher’s LSD, a = 0.05, for each bar, n = 84 clipped leaves and n = 24 leaf litter, 1 S.E.).
p < 0.0001, Fig. 5). Nitrogen content in clipped leaves was highest in the slight decline class, and decreased at both of the more degraded classes, while the nitrogen content of leaf litter was similar at all decline classes (decline class leaf type, F 1,28.5 = 4.53, p < 0.0196, Fig. 6). Clipped water tupelo leaves were higher in nitrogen content than clipped baldcypress leaves, but baldcypress had a higher nitrogen content in leaf litter than water tupelo (species leaf type, F 1,28.6 = 29.61, p < 0.0001, Fig. 7). Foliar nitrogen content significantly increased as percent defoliation increased (covariable in the model—F 1,282 = 38.82, p < 0.0001). Furthermore, simple linear regressions of percent defoliation and foliar nitrogen
301
Fig. 7. Nitrogen content of spring clipped leaves (white) and leaf litter (patterned) of baldcypress and water tupelo trees in 2002 and 2003 (Fisher’s LSD, a = 0.05, for each bar, clipped n = 126, and n = 36 for litter, 1 S.E.).
content showed a significant increase in baldcypress foliar nitrogen in both spring clipped leaves and leaf litter as defoliation increased (Fig. 8). Regressions of tupelo foliage were removed from the analysis based upon violations of normality and homogeneity. 4.4. Comparison of the nitrogen content of frass to other leaf parameters 2002 When all decline classes were analyzed together (2002), a significant interaction occurred between species and leaf types (F 2,204 = 29.17, p < 0.0001, Fig. 9). For water tupelo, frass nitrogen content was significantly higher than clipped leaves and leaf litter. The opposite was true for baldcypress frass, where the nitrogen content was lower than in clipped leaves and leaf litter. Another significant interaction occurred between decline class and leaf type (F 4,204 = 3.85, p = 0.0049, Fig. 10). In the slight decline class, nitrogen content of frass was lower than clipped leaves, but higher than in leaf litter. In the other decline classes, frass nitrogen content was similar in clipped leaves and higher than in leaf litter. Foliar nitrogen content for each tree species and leaf type is summarized in Table 3. 5. Discussion
Fig. 6. Foliar nitrogen content of spring clipped leaves (white) and leaf litter (patterned) of baldcypress and water tupelo trees across three forest conditions in 2002 and 2003 (a priori comparison of high vs. sparse using Fisher’s LSD, a = 0.05, for each bar, clipped n = 84 and n = 24 for litter, 1 S.E.).
Increased basal area growth of both tree species with fertilizer application indicates that moderately stressed areas surrounding Lake Maurepas should benefit most from nutrient enhancement associated with a river diversion. Even severely declining swamp, growth of baldcypress trees should increase. However, water tupelo trees, many of which have severely degraded canopies, are not likely to grow in response to nutrient influxes. In the slight decline class, competition with neighboring tree species may limit basal area growth, masking the effects of
302
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304
Fig. 8. Nitrogen content of clipped leaves (left) and leaf litter (right) in 2002 and 2003 correlated with individual defoliation of baldcypress trees (95% CI).
nutrient augmentation. Increased nutrients may have benefited trees in other ways not detectable by aboveground productivity, for example, through increased nutrient storage, photosynthetic rates, belowground biomass, etc. (Webb, 1978). Although traditional studies often address defoliators as pests, recent studies indicate that arthropods affect forest ecosystems in complex ways. Trees are capable of compensatory growth following defoliator outbreaks, often more than replacing growth loss during defoliation (Alfaro and MacDonald, 1988; Alfaro and Shepherd, 1991; Lovett and Tobiessen, 1993; Trumble et al., 1993). Schowalter et al. (1991) found that defoliators removing 20% of the foliage from young Douglasfir doubled the amount of precipitation reaching the forest floor and increased N, K, and Ca flow to the forest floor by 20–30%. Schowalter et al. (1991) reported that three taxa of litter arthropods were more abundant under defoliated trees, suggesting that defoliation could also increase species diversity and stimulate litter decomposition. Similarly, Progar et al. (2000), Schowalter (1992), and Schowalter et al. (1992, 1998)
demonstrated that saprophytic insects and fungi, including bark beetles and some pathogenic fungi, contribute significantly to nutrient cycling from decomposing wood, and enhance the fertility of underlying soil. Changes in detrital processing and arthropod abundance as a result of defoliation in other forested systems implies that defoliation may be important in forested wetlands in Louisiana affecting, detrital processing and aquatic food webs. The results of this study in Lake Maurepas Swamps ascertained that leaf litter biomass was lowered by defoliation, but defoliation increased overall biomass of litterfall reaching the forest floor when insect frass from the spring was included, particularly for baldcypress. Increased litter reaching the forest floor as a result of defoliation could be particularly important in nutrient limited systems, like Maurepas (Lane et al., 2003). Furthermore, these results indicate that individual baldcypress trees contribute more to the detrital production than water tupelo trees, suggesting that baldcypress may be more influential in nutrient cycling processes.
Fig. 9. Nitrogen content of frass (white), clipped leaves (patterned), and leaf litter (black) for baldcypress and water tupelo trees in 2002 only (Fisher’s LSD, a = 0.05, 1 S.E.).
Fig. 10. Nitrogen content of frass (white), clipped leaves (patterned), and leaf litter (black) for both baldcypress and water tupelo trees at the three decline classes in 2002 only (Fisher’s LSD, a = 0.05, 1 S.E.).
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304 Table 3 Overall mean (1 S.E.) percent nitrogen in frass, clipped leaves, and leaf litter (fallen leaves only) associated with baldcypress and water tupelo trees from 2001 to 2003 in the Lake Maurepas basin Variables
Means
(1 S.E.)
Nitrogen (%) Frass (2001–2003) Clipped leaves (2002–2003) Leaf litter (2002–2003)
Baldcypress 1.171 0.026 1.524 0.032 1.353 0.042
Water tupelo 1.757 0.033 1.632 0.033 1.120 0.042
Fertilizer applications increased foliar nitrogen levels across all forest decline classes, both tree species, and leaf types suggesting that the Lake Maurepas Swamp is nitrogen limited in a wide array of forest decline classes. Both predominant tree species therefore have the capacity to store more organic nitrogen in the form of litter if these swamps are nutrient enhanced (via a diversion), with the N loading rate between of 229–458 m3/s diversion rate. Further, a positive correlation between nitrogen levels in clipped leaves and forest health further supports the hypothesis that nitrogen is a limiting factor in this system. Nitrogen was higher in baldcypress leaf litter than water tupelo leaf litter, suggesting that baldcypress might be a better source of organic nutrients for aquatic food webs and other nutrient-cycling ecosystem processes, depending on overall biomass and decomposition rates. Nitrogen content in water tupelo frass was higher than in baldcypress frass, which could result in this being an important source of nitrogen input during the spring. However, inputs from both frass and leaf litter must be considered in context with total biomass contributions. If BCLR produces a larger quantity of frass than FTC, as suggested from the results, total overall nitrogen loading could be higher during years of peak defoliation. Interestingly, defoliation was positively correlated to nitrogen content in leaf litter, suggesting that defoliators not only increase organic nitrogen loading rates in the spring, but could be causing changes in foliar chemistry that direct a higher nitrogen storage in the leaf litter, as opposed to utilizing resources for other storage organs and biochemical pathways (reproductive structures, roots, wood, defensive compounds, photosynthetic rates, etc.). Average nitrogen levels in clipped leaves for baldcypress in this study were 1.52% 0.032, which is comparable to baldcypress foliage in other swamps at 1.43% (Schlesinger, 1978) and 1.52–1.71% (Meeker, 1992). The increased nutrient levels found in the canopy foliage following fertilization suggest that defoliator populations could potentially be increased. By elucidating patterns of nutrient uptake into leaf tissue and frass (by products), this study has provided results explaining nutrient dynamics not previously addressed for forested wetlands. Furthermore, this information can be applied to the ecosystem as a whole to evaluate the potential of various nutrient loading rates based on forest decline, tree species composition, and defoliation rates (e.g. system models). Scientists studying wetland restoration, especially in the Lake Maurepas and Pontchartrain Basins, will be able to use this information for water quality, fisheries, migratory birds, and
303
other wildlife work in coastal Louisiana and in other wetland ecosystems, and will be able to more accurately predict and assess forested wetland processes affected by variation in nutrients and, often, concurrent insect herbivory. Acknowledgements We would like to thank U.S. Forest Service, Region-8 Forest Health, for providing funding to accomplish this research and LSU Ag Chemistry for conducting nutrient analyses. Others that helped were Ankit Modi, Jay Curole, Wood Johnson, Tessa Bauman, Brendan Watson, Anil Surathu, Brandi Johnson, Chandra Dee Glass (all LSU), and Gary Shaffer (Southeastern Louisiana University). We thank Jay Geaghan (LSU) for assistance in statistical analyses; and Seth Johnson (LSU), Dorothy Prowell (LSU), Michael Stout (LSU), James R. Meeker (U.S. Forest Service), and an anonymous reviewer for their thoughtful comments on earlier versions of this manuscript. This manuscript was approved by the Director of the Louisiana Agricultural Experiment Station as manuscript number 06-26-0133. References Abrahamson, L.P., Harper J.D., 1973. Microbial insecticides control forest tent caterpillar in southwestern Alabama. For. Serv. Res. Note (US) SO-157, p. 3. AOAC, 1995. Official Methods of Analysis of the AOAC, 16th ed. Alfaro, R.I., MacDonald, R.N., 1988. Effects of defoliation by the western false hemloc looper on Douglas-fir tree-ring chronologies. Tree-Ring Bull. 48, 3– 11. Alfaro, R.I., Shepherd, R.F., 1991. Tree-ring growth of interior Douglas-fir after one year’s defoliation by Douglas-fir tussock moth. For. Sci. 37, 959–964. Allen, J.A., Chambers, J.L., Pezeshki, S.R., 1997. Effects of salinity on baldcypress seedlings: physiological responses and their relation to salinity tolerance. Wetlands 17 (2), 310–320. Baumann, R.J., Day, J.W., Miller, C.A., 1984. Mississippi deltaic wetland survival: sedimentation versus coastal survival. Science 224, 1093–1095. Boesch, D.F., Josselyn, M.N., Mehta, A.J., Morris, J.T., Nuttle, W.K., Simenstad, C.A., Swift, D.J.P., 1994. Scientific assessment of coastal wetland loss, restoration and management in Louisiana. J. Coastal Res. (Special issue no. 20), 20. Bormann, F.H., Likens, G.E., 1979. Pattern and Process in a Forested Ecosystem. Springer-Verlag, NewYork, NY, p. 253. Braun, D.M., Goyer, R.A., Lenhard, G.J., 1989. Biology and mortality agents of the fruit tree leafroller (Lepidoptera:Tortricidae), on baldcypress in Louisiana. J. Entomol. Sci 25, 176–184. Brown, C.A., Montz, G.N., 1986. Baldcypress and the Tree Unique, the Wood Eternal. Claitor’s Publishing Division, Baton Rouge, LA, USA, p. 139. Christenson, L.M., Lovett, G.M., Mitchell, M.J., Groffman, P.M., 2002. The fate of nitrogen in gypsy moth frass deposited to an oak forest floor. Oecologia 131 (3), 444–452. Conner, W.H., Brody, M., 1989. Rising water levels and the future of southeastern Louisiana swamps. Estuaries 12, 318–323. Conner, W.H., Buford, M.A., 1998. Southern deepwater swamps. In: Messina, M.G., Conner, W.H. (Eds.), Southern Forested Wetlands: Ecology and Management. Lewis Publishers, NewYork, NY, pp. 261–287. Conner, W.H., Day, J.W., Jr., 1984. The impacts of increased flooding on commercial wetland forests in the Lake Verret watershed. Final Report of Project 83-LBR/022-B to Louisiana Board of Regents. Baton Rouge, LA, USA. Conner, W.H., Day Jr., J.W., 1988. Rising water levels in coastal Louisiana: implications for two coastal forested wetland areas. J. Coastal Res. 4, 589– 596.
304
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304
Conner, W.H., Toliver, J.R., 1990. Long-term trends in the baldcypress (Taxodium distichum) resource in Louisiana (USA). For. Ecol. Manage. 33–34, 543–557. Coulson, R.N., Witter, J.A., 1984. Forest Entomology: Ecology and Management. Wiley Publishing Co., NY, p. 699. Day Jr., J.W., Shaffer, G.P., Britsch, L.D., Reed, D.J., Hawes, S.R., Cahoon, D., 2000. Pattern and process of land loss in the Louisiana coastal zone: an analysis of spatial and temporal patterns of wetland habitat change. Estuaries 23, 425–438. Day Jr., J.W., Arancibia, A.Y., Mitsch, W.J., Lara-Dominguez, A.L., Day, J.N., Ko, J.Y., Lane, R., Linsey, J., Lomeli, D.A., 2003. Using ecotechnology to address water quality in wetland habitat loss problems in the Mississippi basin: a hierarchical approach. Biotechnol. Adv. 22, 135–159. Day Jr., J.W., Conner, W.H., Shaffer, G.P., 2006. The importance of pulsed physical events for sustainability of Louisiana coastal forested wetlands. In: Nettles, J., Williams, T. (Eds.), Hydrol. Manage. Forested Wetlands Proc. Int. Conf., American Society of Agricultural and Biological Engineers, St. Joseph, MI, pp. 156–158. Dennis, J.V., 1988. The Great Cypress Swamps. Louisiana State University Press, Baton Rouge, LA, USA, p. 142. Goyer, R.A., Chambers, J., 1996. Evolution of insect defoliation in baldcypress and its relationship to flooding. USDI. NBS. Biol. Sci. Rep. 8. Goyer, R.A., Lenhard, G.J., 1988. A new insect pest threatens baldcypress. LA Agriculture, LA Agricultural Experiment Station 31, pp. 16–17, 21. Hutchens, J.J., Benfield, E.F., 2000. Effects of forest defoliation by the gypsy moth on detritus processing in southern Appalachian streams. Am. Midland Nat. 143 (2), 397–404. Johnson, C.W., 2004. The performance of the baldcypress leafroller (Archips goyerana Kruse, Lepidoptera:Tortricidae) in response to fertilization, thinning, and genetic variation in host baldcypress (Taxodium distichum L. Richard). MS Thesis. Louisiana State University, Baton Rouge, LA. Kemp, P.G., Conner, W.H., Day Jr., J.W., 1985. Effects of flooding on decomposition and nutrient cycling in a Louisiana swamp forest. Wetlands 5, 35– 50. Klock, G.O., Wickman, B.E., 1978. Ecosystems effects. In: Brookes, M.H., Stark, R.W., Campbell, R.W. (Eds.), The Douglas-Fir Tussock Moth: A Synthesis. USDA For. Ser. Technical Bull. 1585, pp. 90–94. Lane, R.R., Day, J.W., Thibodeaux, B., 1999. Water quality analysis of a freshwater diversion at Caernarvon, Louisiana. Estuaries 22, 327– 336. Lane, R.R., Mashriqui, H.S., Kemp, G.P., Day, J.W., Day, J.N., Hamilton, A., 2003. Potential nitrate removal from a river diversion into a Mississippi delta forested wetland. Ecol. Eng. 20, 237–249. Louisiana Coastal Wetlands Conservation and Restoration Task Force, 1993. Louisiana Coastal Wetlands Plan. Louisiana Wetlands Coastal Planning. Protection, and Restoration Act, Baton Rouge, LA, USA. Louisiana Coastal Wetlands Conservation and Restoration Task Force and the Wetlands Conservation and Restoration Authority, 1998. Coast 2050: Toward a Sustainable Louisiana. Louisiana Department of Natural Recources, Baton Rouge, LA, USA. Lovett, G., Tobiessen, P., 1993. Carbon and nitrogen assimilation in red oaks (Quercus rubra L.) subject to defoliation and nitrogen stress. Tree Physiol. 12, 259–269. Mancil, E., 1980. Pullboat logging. J. For. Hist. 24, 135–141. Meeker, J.R., 1992. Host quality of baldcypress and its influence of the fruit tree leafroller, Archips agyrospila (Walker) (Lepidoptera: Tortricidae), Performance in Forested Wetlands of Louisiana. A Dissertation. Louisiana State University. Baton Rouge, LA, USA. Mitsch, W.J., Gosselink, J.G., 1993. Wetlands, second ed. Van Norstrand Reinhold, New York, NY. Mossa, J., 1996. Sediment dynamics in the lowermost Mississippi River. Eng. Geol. 45, 457–479. Nachod, L.H., 1977. Spring defoliation by forest insects in Louisiana. Insect and Disease Report. Louisiana Office of Forestry, Woodworth, LA, USA, p. 2. Nachod, L.H., Kucera, D.R., 1971. Observations of the forest tent caterpillar in south Louisiana. Insect and Disease Report. Louisiana Office of Forestry, Woodworth, LA, USA, p. 2.
Progar, R.A., Schowalter, T.D., Morrell, J.J., Freitag, C.M., 2000. Respiration from coarse woody debris as affected by moisture and saprotroph functional diversity in western Oregon. Oecologia 124, 426–431. Salinas, L.M., DeLaune, R.D., Patrick Jr., W.H., 1986. Changes occurring along a rapidly emerging coastal area: Louisiana, USA. J. Coastal Res. 2, 269– 284. SAS Institute, Inc., 1999. SAS/STAT User’s Guide Version 8.0. Cary, NC. Schlesinger, W.H., 1978. Community structure, dynamics and nutrient cycling in the Okefenokee cypress swamp-forest. Ecol. Monogr. 48, 43–65. Schowalter, T.D., 1981. Insect herbivore relationship to the state of the host plant: biotic relation of ecosystem nutrient cycling through ecosystem succession. Oikos 37, 126–130. Schowalter, T.D., 1992. Early decomposition and nutrient dynamics of oak (Quercus) logs at four sites across a North American gradient. Can. J. For. Res. 22, 161–166. Schowalter, T.D., Caldwell, B.A., Carpenter, S.E., Griffiths, R.P., Harmon, M.E., Ingham, E.R., Kelsey, R.G., Lattin, J.D., Moldenke, A.R., 1992. Decomposition of fallen trees: effects of initial conditions and heterotroph colonization rate. In: Singh, K.P. (Ed.), Tropical Ecosystems: Ecology and Management. Wiley Eastern, Ltd., New Delhi, pp. 371–381. Schowalter, T.D., Crossley Jr., D.A., 1983. Forest canopy arthropods as sodium, potassium, magnesium and calcium pools in forests. For. Ecol. Manage. 7, 143–148. Schowalter, T.D., Sabin, T.E., Stafford, S.G., Sexton, J.M., 1991. Phytophage effects primary production, nutrient turnover, and litter decomposition on young Douglas-fir in western Oregon. For. Ecol. Manage. 42, 229–243. Schowalter, T.D., Zhang, Y.L., Sabin, T.E., 1998. Decomposition and nutrient dynamics of oak (Quercus spp.) logs after five years of decomposition. Ecography 21, 3–10. Seastedt, T.R., Crossley Jr., D.A., 1984. The influence of arthropods on ecosystems. BioScience 34, 157–161. Shaffer, G.P., Perkins, T.E., Hoeppner, S.S., Howell, S., Beneard, T.H., Parsons, A.C., 2003. Ecosystem health of the Maurepas Swamp: feasibility and projected benefits of a freshwater diversion. Final Report. Environmental Protection Agency, Dallas, TX, Region 6, p. 105. Souther-Effler, R.F., 2004. Interactions of insect herbivory and multiple abiotic stress agents on two wetland tree species in southeast Louisiana swamps. PhD Dissertation. Louisiana State University, Baton Rouge, LA, USA. Swank, W.T., Waide, J.B., Crossley Jr., D.A., Todd, R.L., 1981. Insect defoliation enhances nitrate export from forest ecosystems. Oecologia 51, 297– 299. Templet, P.H., Meyer-Arendt, K.J., 1988. Louisiana wetland loss: regional water management approach to the problem. Environ. Manage. 12, 181– 192. Trumble, J.T., Kolodny-Hirsch, D.M., Ting, I.P., 1993. Plant compensation for arthropod herbivory. Annu. Rev. Entomol. 38, 93–119. Turner, R.E., 1997. Wetland loss in the northern Gulf of Mexico: multiple working hypotheses. Estuaries 20, 1–13. Thompson, D.A., 2000. The influence of Hydrological alterations upon wetland hydrodynamics and plant growth on the Manchac Landbridge, Southeastern Louisiana, USA. Master’s Southeastern Louisiana University, Hammond, LA. Visser, J.M., 1989. The impact of vertebrate herbivores on primary production of Sagittaria marshes in the Wax Lake Delta, Atchafalaya Bay, Louisiana. PhD Dissertation. Louisiana State University, Baton Rouge, LA, USA. Webb, W.L., 1978. Effects of defoliation and tree energetics. In: Brookes, M.H., Stark, R.W., Campbell, R.W. (Eds.), The Douglas-Fir Tussock Moth: A Synthesis. USDA For. Serv. Technical Bull. 1585, pp. 77–81. White, W.B., Schneeberger, N.F., 1981. Socioeconomic impacts. In: Doane, C.C., McManus, M.L. (Eds.). The Gypsy Moth: Research Toward Integrated Pest Management. USDA For. Ser. Technical Bull. 1584, pp. 681–694. Wilson, L., and Associates, Shaffer, G., Hester, M., Kemp, P., Mashriqui, H., Day, J., Lane, R., 2001. Diversion into Maurepas Swamp. Project Report to U.S. Environmental Protection Agency, Region 6, Dallas, TX Contract No. 68-06-0067, WA #5-02, June 2001. Zlotin, R.I., Khodashova, K.S., 1980. In: French, N.R. (Ed.), The Role of Animals in Biological Cycling of Forest-Steppe Ecosystems (English translation). Dowden, Hutchinson, and Ross, Stroudsburg, PA, p. 221.
Baldcypress and water tupelo responses to insect defoliation and nutrient augmentation in Maurepas Swamp, Louisiana, USA Rebecca S. Effler *, Richard A. Goyer, Gerald J. Lenhard LSU AgCenter, Department of Entomology, Louisiana State University, 404 Life Sciences, Baton Rouge, LA 70803, USA Received 28 February 2006; received in revised form 12 September 2006; accepted 13 September 2006
Abstract Forested wetlands in Louisiana are hydrologically isolated from the Mississippi River, impacted by saltwater intrusion, and are sinking, resulting in more frequent flooding for longer periods. These problems are exacerbated in some areas by impoundment from man-made structures. Additionally, defoliation of the two dominant trees, baldcypress (Taxodium distichum L. Rich) and water tupelo (Nyssa aquatica L.), frequently occur during spring. In Louisiana, the baldcypress leafroller, BCLR (Archips goyerana Kruse) and the forest tent caterpillar, FTC (Malacosoma disstria Hubner) annually defoliate an average of 120,000 ha of baldcypress–tupelo swamps. In combination, these factors have led to severely degraded forested wetlands with reduced basal area, dieback and tree mortality over extensive, formerly productive swamplands. Restoration plans call for reintroducing Mississippi River water by diversions, to increase sediment elevation, promote natural regeneration, and enhance primary and secondary productivity. Nutrient augmentation of baldcypress and water tupelo trees in Maurepas Swamps (similar to NO3 loading rates expected from 229 m3/s diversion) increased radial growth (cm2) of both species (especially baldcypress) in degraded forest stands. Nutrient augmentation also increased nitrogen in foliage (insect frass (feces), spring clipped leaves, and abscised leaf litter) for both tree species. Nitrogen content of canopy foliage and leaf litter was positively correlated with site health. These findings support hypotheses that swamps in southeastern Louisiana are nutrient limited, and existing trees can utilize, benefit, and act as nutrient sinks for nutrient-laden river water accompanying diversions. Defoliation of these swamp species, although not uniform among years, did not result in losses in radial growth, but increased overall litter deposition for baldcypress. An increase in nitrogen content in insect frass in response to nutrient augmentation indicates defoliator populations might increase as a result of nutrient influxes. # 2006 Elsevier B.V. All rights reserved. Keywords: Diversions; Taxodium distichum; Nyssa aquatica; Malacosoma disstria; Archips goyerana; Lepidoptera; Nitrogen; Litterfall
1. Introduction Forested wetlands are a valuable natural resource that stimulates and sustains both economic productivity and ecological functioning on local and global scales. They have been identified as important nutrient sinks that enhance water quality and export organic debris to many aquatic food webs (Mitsch and Gosselink, 1993). However forested wetlands are being lost worldwide due mainly to altered natural hydrology. In Louisiana, USA, freshwater input from the largest North American river system (Mississippi River) has been reduced in or in some cases completely isolated from forested wetlands by
* Corresponding author. Present address: University of Georgia Marine Institute, Sapelo Island, GA 31327, USA. Tel.: +1 912 485 2225; fax: +1 912 485 2133. E-mail address: [email protected] (R.S. Effler). 0378-1127/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.foreco.2006.09.014
levees and dams (Louisiana Coastal Wetlands Conservation, 1998). Wetlands that were historically influenced, but now isolated from these riverine inputs, are currently subsiding at approximately 1 m/century (Salinas et al., 1986), and decreasing in productivity and biodiversity (Conner and Day, 1988; Conner and Buford, 1998). Many swamps in coastal Louisiana, including Maurepas Basin, are hydrologically isolated from the Mississippi River (Mossa, 1996), predominantly impounded by man-made structures (Turner, 1997), degraded by saltwater intrusion (Conner and Brody, 1989; Conner and Toliver, 1990), and have experienced more frequent flooding, for longer periods (Conner and Day, 1984). To sustain current canopy trees and promote natural regeneration, Mississippi River inputs (nutrient flows and other riverine inputs by diversions) need to be re-established into declining wetland systems (Baumann et al., 1984; Templet and Meyer-Arendt, 1988; Boesch et al., 1994; Day et al., 2000, 2003). Active and
296
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304
approved diversions are aimed to flush out soil phytotoxins and interstitial salt accumulation from saltwater events, and increase elevation and production as historically accomplished by natural hydrological conditions (Louisiana Coastal Wetlands Conservation, 1993, 1998). Two diversions in operation (Atchafalaya Spillway and Caernarvon) have demonstrated that riverine inputs into marshes result in increased primary and secondary production (Visser, 1989; Louisiana Coastal Wetlands Conservation, 1998; Lane et al., 1999). In addition to abiotic stressors present (impoundment, subsidence, increased salinity, nutrient deprivation) in southeast Louisiana swamps, the two dominant tree species, water tupelo (Nyssa aquatica L.) and baldcypress (Taxodium distichum L. Richard), are subject to frequent springtime defoliation. This defoliation event simultaneously occurs for the two species and starts in March and continues through mid May. Water tupelo is defoliated by the forest tent caterpillar, FTC (Malacosoma disstria Hubner). Defoliation levels in south Louisiana alone have exceeded 250,000 ha during a single season (Nachod and Kucera, 1971; Nachod, 1977) and average 120,000 ha each year with much occurring in Maurepas Basin (Goyer and Chambers, 1996). Abrahamson and Harper (1973) found that up to 45% of the annual incremental growth of water tupelo was lost due to severe FTC defoliation during a 5-year impact study in Alabama. In 1983, an epidemic of the baldcypress leafroller, BCLR (Archips goyerana Kruse) occurred on baldcypress in and around the Atchafalaya Basin (Goyer and Lenhard, 1988). This outbreak has since spread to southern Lake Maurepas and Lake Verret (Goyer and Chambers, 1996) defoliating as much as 200,000 ha in a single season and averaging 72,000 ha each year. Baldcypress, renowned for its lack of serious insect and disease problems (Brown and Montz, 1986), presently experiences defoliation damage resulting in significant reduction in radial growth and dieback (Braun et al., 1989; Allen et al., 1997; Goyer and Chambers, 1996). Baldcypress health generates concern among biologists because it is one of the most ecologically valuable tree species in the southeastern floodplain forests (Brown and Montz, 1986). Baldcypress maintains forest ecosystem integrity through its extensive root system, canopy, and long-lived nature. It is the keystone habitat species for many game animals, sport fish, and formerly or presently threatened and endangered species such as the bald eagle (Haliaeetus leucophalus (L.)), Louisiana black bear (Ursus americanus luteolus Griffith), American alligator (Alligator mississippiensis (Daudin)), Florida cougar (Felis concolor coryi (Bangs)) (Dennis, 1988), and potentially the recently rediscovered ivory-billed woodpecker. Forest defoliation by insects can also lead to severe disruptions in nutrient cycling by altering biomass, detrital processes, and nutrient content of litterfall reaching the forest floor. Studies have shown that defoliation products (foliage, insect tissues, and frass) transfer a considerable amount of nutrients to the litter layer (Schowalter, 1981; Schowalter and Crossley, 1983; Seastedt and Crossley, 1984; Swank et al., 1981). Premature litterfall resulting from artificial defoliation returned 10 times as much nitrogen to the litter as would normal
litterfall (Klock and Wickman, 1978). Schowalter et al. (1991) found that defoliators removing 20% of the foliage from young Douglas-fir (Pseudotsuga menseii (Miirb.) Franco), doubled the amount of precipitation reaching the forest floor, and increased N, K, and Ca flow to the forest floor by 20–30%. With more nutrients reaching the forest floor, defoliation may affect water quality in adjacent streams (Coulson and Witter, 1984; Klock and Wickman, 1978; Swank et al., 1981; White and Schneeberger, 1981). In the northeastern United States, herbivory does not appear to significantly affect water quality (Bormann and Likens, 1979; White and Schneeberger, 1981), perhaps because of the high baseline concentration of nutrients in these streams (Swank et al., 1981). However, nitrateimpoverished streams in the southeastern United States show significant increases in NO3-N concentrations during periods of defoliation (Swank et al., 1981). Hutchens and Benfield (2000) found that breakdown rates of chestnut oak (Quercus prinus L.) leaves were faster for secondflush leaves produced after gypsy moth (Lymantria dispar L.) defoliation than those of natural spring-flush leaves shed in autumn in three of six streams tested. They concluded that insect defoliation accelerates detritus processing in southern Appalachian streams. However, induced defenses by the plant as a result of insect pressure may reduce litter decomposition rates (Zlotin and Khodashova, 1980). Based on N-15 labeled isotopes, N-15 in gypsy moth frass was mobilized more quickly (40% incorporated in the soil) than N-15 in leaf litter (80% remained undecomposed in the leaf litter) in an oak forest (Christenson et al., 2002). However, frass N may be largely unavailable to plants and microorganisms, as little N-15 was found in extractable, microbial, or readily mineralizable pools. These processes may be particularly important for declining, nutrient-limited Louisiana swamps, as litterfall associated with insect herbivory may be an important nutrient source for primary and secondary productivity in the spring when biological emergence is at its peak. Importantly, baldcypress–tupelo swamps are known to be long-term sinks for both nitrogen and phosphorous via burial of partially decomposed organic and inorganic matter under reduced soil conditions and denitrification (Kemp et al., 1985). However, little research has been conducted into nutrient/biological interactions of the trees and defoliators in these swamps, with the exception of Meeker (1992) and Johnson (2004) that address quality of baldcypress foliage and insect performance. During BCLR larval activity, nutrient concentration (N, P, K, and Mg) of baldcypress foliage decreased as feeding progressed; and constituents representing tree defense mechanisms increased (Meeker, 1992), rapidly creating foliage characteristic unfavorable for the BCLR. Meeker (1992) also found that a flooded swamp may have more suitable foliage for the larvae, as suggested by having higher total nonstructural carbohydrates in foliage than in an unflooded swamp. Objectives of this study were to determine if nutrient enhancement of mature baldcypress and water tupelo trees will likely (1) increase stem growth, litter biomass, and foliar nitrogen content of trees in swamps exhibiting various states of decline, and (2) compensate for basal area growth loss as a result of defoliation by insects. Also, (3) determine the
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304
influence of defoliation on annual biomass of litterfall of mature baldcypress and water tupelo trees in these areas. 2. Methods 2.1. Description of Maurepas Swamps Maurepas Swamps are located in the Lake Pontchartrain Basin in southeast Louisiana (Fig. 1). Lake Maurepas (308100 N and 918300 W) is a large lake roughly 21 km in diameter. At the turn of the 20th Century, swamps in this basin, as well as most of coastal Louisiana, were harvested en masse (Mancil, 1980). Maurepas forested wetlands are the second largest contiguous coastal forest and consist of 77,550 ha of swamps and 5204 ha of fresh/intermediate marshes (LACOAST 2050, 1998). Swamps along the southern shore of Lake Maurepas are composed of 80% baldcypress and 20% water tupelo trees (Wilson et al., 2001), while deeper areas west of the lake show the reverse composition. The key driving factors of forest decline are saltwater intrusion, land subsidence, and lack of nutrient inputs (Thompson, 2000; Wilson et al., 2001; Lane et al., 2003; Shaffer et al., 2003). Insect defoliation in addition to these stressor further declines baldcypress and water tupelo health. The BCLR has recently reached outbreak proportions in southern Maurepas Swamps (1996–1999). Populations of FTC have been locally epidemic throughout the swamps since 1948, when records were kept. This swamp system is a candidate
297
hydrological restoration area (via river diversions) because of its rapid loss of swamp habitat and its previous history of receiving inputs from the Mississippi River. Furthermore, most nutrients accompanying the proposed diversion for the area are expected to be retained in the swamp and should not adversely affect water quality in Lake Maurepas (Lane et al., 2003). 2.2. Experimental design The Lake Maurepas field study to evaluate effects of nutrient augmentation on tree above-ground productivity was conducted from December 2000 to December 2003. The experiment consisted of a split-plot design. The main plot included three forest decline classes were (slight, moderate and severe). Locations of these decline classes are located in Fig. 1. Within each decline class, two plots were installed. Each plot consisted of a 2 3 factorial design. Within each decline class, two tree species (baldcypress and water tupelo) were assessed. Each tree species randomly received three nutrient regimes [no nutrients, a 1 dose of 236 g/m2 of Osmocote1 18-6-12 (B.W.I. Jackson, MS, USA), and a 2 dose]. The 1 rate is equivalent to a N loading rate of 229 m3/s springtime effluent of a Mississippi River diversion (Lane et al., 1999; Day et al., 2006). Each decline class had seven replicates (trees) of each fertilizer-species combination totaling 42 trees per decline class. The seven replicates were divided between the two plots in each decline class resulting in either three or
Fig. 1. Location of Maurepas Swamp and study sites.
298
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304
four replicate trees per plot. Trees were estimated to be between 40 and 80 years old. Trees were selected based upon minimum overlapping canopies to allow for litterfall collection. Osmocote fertilizer was applied each spring from 2001 to 2003 prior to bud break. Area of coverage was determined by estimating average canopy radius (dripline) to the nearest 0.5 m. The fertilizer was uniformly distributed equidistant around the base of the tree once a circular area was established. All areas selected also had recent defoliator activity on baldcypress and water tupelo. 2.3. Data collection Aluminum growth bands (1.5 cm 0.1 cm) were placed on all trees in January 2001 at breast height or at 0.5 m above butt swell on larger trees. Stem growth was monitored from January 2001 to December 2003 and recorded each year during dormancy (January), after defoliation (May), and late-summer (September) to the nearest 0.025 cm. A single litterfall/frass collection trap was placed under each tree. Litterfall traps were positioned 1 m above the forest floor and consisted of a wooden frame constructed from #2 treated lumber with a surface area of 0.25 m2 and 10 cm in depth. Trap bases were lined with 2 mm vinyl screen. Litterfall was collected once every 2 months, dried, and separated by plant species, and leaf material was weighed to the nearest 0.001 g. After refoliation occurred in June, leaves from the crown on the trees were sampled from approximately 10 to 15 m height by pole pruner or shotgun (clipped leaves). Subsamples of both clipped (spring) and leaf litter (natural abscission) were analyzed for N content. During spring (peak feeding activity of both insect species), these traps served as frass (feces) collection traps by inserting fine, saran screening (0.25 mm mesh) stretched over a 1.9 cm (3/4 in.) PVC pipe frame constructed to fit tightly (approximately the same area) inside the litterfall trap. Frass was collected once a week by sweeping the frass trap into a 120 ml vinyl specimen cup. Samples were returned to lab, uncovered, oven dried at 65 8C for 48 h, and then stored at room temperature until they could be separated from other litter and weighed to the nearest 0.001 g. During the last week of defoliation, percent defoliation of each tree was estimated to the nearest 10%. To more accurately measure annual leaf productivity across all defoliation levels, ‘‘frass’’ biomass was added to end of the year ‘‘leaf litter’’ (=‘‘litterfall’’) of each species and trends between the two were compared. In 2002 and 2003, subsamples of both clipped leaves and leaf litter were analyzed for N–P–K content (see SoutherEffler, 2004). For clarity, and its importance to both insect and tree nutrition, only N content is reported here. Nitrogen analyses were conducted on subsamples of frass (from peak collection period for each species) to evaluate nutrient inputs from insect herbivory independently from undefoliated, senesced leaves, which presumably contained a lower nutrient composition than frass. The 2002 frass was the only year that had representation of all treatment combinations. Thus, a second analysis was conducted in order to directly compare nutrient content of frass to litter and foliar parameters.
Nitrogen content of all foliage and frass was ascertained using AOAC (1995) protocols at the LSU/LDAF Agricultural Chemistry Laboratory on the LSU-Baton Rouge campus. 3. Statistical analyses Initially, all variables were analyzed as a split-plot design, as described in Section 2. However, plot variation in each decline class did not statistically differ (non-significant F-value < 1.7), so the degrees of freedom and sum of squares were pooled. Therefore, unless specifically stated, all variables were analyzed as a completely randomized design (CRD) with a 2 3 3 factorial arrangement: two species (baldcypress and water tupelo), three fertilizer treatments (0, 1, and 2), and three decline classes (slight, moderate, and severe decline). Fisher’s LSD was used for specific comparisons of interaction components. 3.1. Basal area growth An ANCOVA was conducted using the proc MIXED statement on SAS 8.0 (SAS Inst. Inc., 1999). The response variable was basal area growth (cm2/2 years) from January 2002 to December 2003. These data were analyzed as a CRD with a 2 3 3 factorial arrangement. To account for larger trees yielding higher basal area growth than smaller trees, initial stem basal area was added as a covariable. Two outliers were removed from this analysis. Further, analyses were conducted to test if fertilization compensated for loss of basal area growth resulting from defoliation. Yearly basal area growth from 2002 and 2003 of each species was regressed with defoliation from that year in each of the fertilizer treatments. Again, stem basal area was added as a covariable in each model. Each of the six models had 42 observations. 3.2. Litter productivity Two ANCOVAs were conducted on litter productivity, one on the biomass of leaf litter (fallen leaves only) and the other on the biomass of litterfall (leaf litter plus frass) (weighed to the nearest 0.001 g/0.25 m2/year). Defoliation (%) of each tree for that year served as a covariable in each model to test for the effects on leaf productivity. Initially, these variables were analyzed as a randomized block design (RBD) (with a block on the year— 2002 and 2003 collections) with a 2 3 3 factorial arrangement. However, variation due to year did not statistically differ for both analyses, so the degrees of freedom and sum of squares were pooled into the overall model. Therefore, the data were analyzed as a CRD with a 2 3 3 factorial arrangement. In addition, simple linear regression equations of the covariable (defoliation) were conducted for both leaf litter and litterfall. 3.3. Nitrogen content of clipped leaves and leaf litter 2002–2003 An ANCOVA was conducted on nitrogen (%) content with defoliation (%) as a covariable. Nitrogen (%) was analyzed as a
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304
randomized block design (RBD) (with a block on the year— 2002 and 2003 collections) with 2 2 3 3 factorial arrangement. The treatments included two leaf types (spring clipped leaves and leaf litter), two species (baldcypress and water tupelo), three fertilizer treatments (0, 1, and 2), and three decline classes (slight, moderate, and severe decline). Each leaf type was represented in each of the cross-classified species, fertility, and decline class categories. The number of trees sampled each year was n = 36 for litter and n = 122 for clipped for 2002, and n = 36 for leaf litter and n = 126 for clipped leaves for 2003. Degrees of freedom were Satterthwaite adjusted for unequal sample sizes. One outlier from the clipped leaves was removed. In addition, simple linear regression equations of the covariable (defoliation) were conducted for both leaf litter and litterfall. 3.4. Comparison of the nitrogen content of frass to other leaf parameters in 2002 As stated in Section 2, the 2002 data set was the only year frass was represented in all treatment combinations, and therefore a model was constructed for that year only, comparing relative amounts of nitrogen in frass to that in clipped leaves and leaf litter. An ANOVA was conducted on nitrogen (%) content as a CRD with a 2 3 3 3 factorial arrangement. The treatments were as above (Section 3.3) except that three leaf types (frass, spring clipped leaves, and leaf litter) were utilized. Each leaf type was represented in each of the cross-classified species, fertility, and decline class categories. Number of trees sampled for each leaf type was n = 101 for frass, n = 122 for clipped, and n = 36 for leaf litter. One outlier from the clipped leaves was removed. 4. Results Specific sites for this study represent three distinct degrees of decline in the Maurepas Basin—slight, moderate, and severe decline, based on a swamp productivity survey by Wilson et al. (2001). In this system, tree density is somewhat reflective of the varying degrees of forest decline. Basal area (BA) for each decline class in this study were slight decline class, BA = 32.84 1.14 m2/ha, moderate decline, BA = 15.72 0.94 m2/ha, severe decline BA = 16.73 0.90 m2/ha. Although basal area density of the severe decline class was similar to that of the moderate decline, the severe decline class had less canopy closure due to crown loss and dieback, especially water tupelo. These trees were missing the apical portion of the main stem resulting in mostly coppiced growth, indicative of the most degraded swamps in the basin where trees still exist. In addition, sites were also required to have recent insect defoliation of both tree species. Percent defoliation and diameter (cm) means and ranges of study trees are summarized in Tables 1 and 2. 4.1. Basal area growth Although the three-way interaction was not significant in the overall model, six a priori comparisons (using Fisher’s LSD) were used to determine if fertilization increased growth for each
299
Table 1 Diameter at breast height (cm) of study trees, by species for the decline class in 2002–2003 (n = 21 for each mean) Decline
Species
Mean (1 S.E.)
Minimum
Maximum
Slight
Water tupelo Baldcypress
24.9 1.0 31.9 1.9
16.5 14.5
34.5 48.3
Moderate
Water tupelo Baldcypress
30.3 1.1 25.2 1.7
21.6 11.7
41.9 46.0
Severe
Water tupelo Baldcypress
36.7 2.2 33.0 2.4
14.5 13.5
56.6 56.4
species in each decline class (Fig. 2). Results show that neither species benefited from fertilization application at the slight decline class; both species grew more with the fertilizer applications in the moderate decline class; and only baldcypress grew more with the fertilizer applications at the severe decline class while water tupelo did not. Basal area growth was highest in the moderate decline class (decline class, F 2,105 = 19.78, p < 0.0001, Fig. 2), which was mainly due to the differential utilization of fertilizer in each decline class (decline class fertilizer, F 4,105 = 3.51, p = 0.0099, Fig. 2). Fertilization in the moderate decline class led to increased tree growth compared to either the slight or severe decline class. Also, baldcypress trees grew more in basal area than water tupelo trees (F 1,105 = 23.99, p < 0.0001, Fig. 2). Defoliation (%) could not be correlated to basal area growth. 4.2. Litter productivity Fertilization did not significantly increase biomass of leaf litter (F 2,233 = 1.38; p = 0.2533). Leaf litter was highest under trees at the slight decline class, followed by the moderate decline class, and lowest at the severe decline class (Fig. 3, gray bars). This trend differed with regard to species; baldcypress trees produced significantly more leaves per tree at the slight Table 2 Percent defoliation of study trees, by species, for the decline class in 2002 and 2003 (n = 21 for each mean) Year/decline
Species
Mean (1 S.E.)
Minimum
Maximum
Water tupelo Baldcypress
99.5 0.5 77.9 4.8
90 30
100 100
Moderate
Water tupelo Baldcypress
12.4 4.6 72.9 6.1
1 15
70 100
Severe
Water tupelo Baldcypress
61.6 9.8 3.6 1.4
1 1
100 30
Water tupelo Baldcypress
99.3 0.5 62.6 6.4
90 10
100 100
Moderate
Water tupelo Baldcypress
2.7 0.9 71.7 6.3
1 20
20 100
Severe
Water tupelo Baldcypress
3.6 1.4 2.2 0.5
1 1
30 10
2002 Slight
2003 Slight
300
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304
Fig. 4. Baldcypress litterfall (frass and leaf litter) (g/0.25 m2/year) and percent defoliation of each tree for 2002 and 2003 (n = 126, 95% CI).
Fig. 2. Cumulative basal area growth (cm2/2 years) of (a) water tupelo and (b) baldcypress trees after receiving three levels of fertilization (1 = 236 g/m2 of Osmocote1 18-6-12) in three forest decline classes from 2002 to 2003 (n = 7 trees for each bar, 1 S.E.).
and severe decline classes, while water tupelo had a higher leaf litter than baldcypress at the moderate decline class (decline class species, F 2,232 = 5.07, p = 0.0070, Fig. 3, gray bars). As defoliation increased, a covariable in the model, biomass of leaf litter decreased (F 1,232 = 4.60, p = 0.0330). When defoliation was regressed by each species by decline class, only baldcypress at the moderate decline class was significant (r2 = 0.323, y = 0.304x + 57.986, F 1,40 = 19.07, p < 0.0001). Results for litterfall (leaf litter plus frass) showed that baldcypress produced significantly more litterfall than water tupelo at all decline classes (F 1,232 = 108.71, p < 0.0001, Fig. 3, white bars). Furthermore, baldcypress trees’ litterfall significantly decreased as forest health decreased; while water tupelo had similar litterfall productivity in the slight and moderate decline classes, but significantly less at the severe decline class (decline class species, F 2,232 = 5.07, p = 0.0070, Fig. 3). In this model, as defoliation increased, a covariable in the model, biomass of litterfall increased (F 1,232 = 4.60, p = 0.0330). Simple linear regressions of percent defoliation and litterfall showed a significant increase in baldcypress litterfall productivity as defoliation increased (Fig. 4), while tupelo did not. This is in contrast to the leaf litter analysis, where defoliation was negatively correlated to leaf litter productivity. 4.3. Nitrogen content of clipped leaves and leaf litter 2002–2003
Fig. 3. Mean annual litterfall (g/0.25 m2/year), leaf litter (leaf fall only – bottom gray) and litterfall (leaf litter + frass top white) of baldcypress (solid) and water tupelo (patterned) trees in three forest decline classes from 2002 to 2003 (n = 42 for each bar, 1 S.E., Fisher’s LSD, a = 0.05 for both leaf litter and litterfall).
As fertilization increased, foliar nitrogen levels significantly increased (main effect—F 2,28.9 = 7.76, p = 0.020, linear contrast—F 1,29 = 15.51, p = 0.0005, Fig. 5). Fertilizer application did not interact with leaf type, species, or decline class, indicating that foliar nitrogen increased as a result of fertilization in both spring clipped and leaf litter samples, in both baldcypress and water tupelo and in trees in each of the forest decline classes. As expected, spring clipped leaves had a higher nitrogen content than those in leaf litter (F 1,29 = 89.06,
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304
Fig. 5. Foliar nitrogen content of spring clipped leaves and leaf litter in response to fertilizer application to baldcypress and water tupelo trees (1 = 236 g/m2 of Osmocote1 18-6-12, a priori comparison of 0 vs. 2 using Fisher’s LSD, a = 0.05, for each bar, n = 84 clipped leaves and n = 24 leaf litter, 1 S.E.).
p < 0.0001, Fig. 5). Nitrogen content in clipped leaves was highest in the slight decline class, and decreased at both of the more degraded classes, while the nitrogen content of leaf litter was similar at all decline classes (decline class leaf type, F 1,28.5 = 4.53, p < 0.0196, Fig. 6). Clipped water tupelo leaves were higher in nitrogen content than clipped baldcypress leaves, but baldcypress had a higher nitrogen content in leaf litter than water tupelo (species leaf type, F 1,28.6 = 29.61, p < 0.0001, Fig. 7). Foliar nitrogen content significantly increased as percent defoliation increased (covariable in the model—F 1,282 = 38.82, p < 0.0001). Furthermore, simple linear regressions of percent defoliation and foliar nitrogen
301
Fig. 7. Nitrogen content of spring clipped leaves (white) and leaf litter (patterned) of baldcypress and water tupelo trees in 2002 and 2003 (Fisher’s LSD, a = 0.05, for each bar, clipped n = 126, and n = 36 for litter, 1 S.E.).
content showed a significant increase in baldcypress foliar nitrogen in both spring clipped leaves and leaf litter as defoliation increased (Fig. 8). Regressions of tupelo foliage were removed from the analysis based upon violations of normality and homogeneity. 4.4. Comparison of the nitrogen content of frass to other leaf parameters 2002 When all decline classes were analyzed together (2002), a significant interaction occurred between species and leaf types (F 2,204 = 29.17, p < 0.0001, Fig. 9). For water tupelo, frass nitrogen content was significantly higher than clipped leaves and leaf litter. The opposite was true for baldcypress frass, where the nitrogen content was lower than in clipped leaves and leaf litter. Another significant interaction occurred between decline class and leaf type (F 4,204 = 3.85, p = 0.0049, Fig. 10). In the slight decline class, nitrogen content of frass was lower than clipped leaves, but higher than in leaf litter. In the other decline classes, frass nitrogen content was similar in clipped leaves and higher than in leaf litter. Foliar nitrogen content for each tree species and leaf type is summarized in Table 3. 5. Discussion
Fig. 6. Foliar nitrogen content of spring clipped leaves (white) and leaf litter (patterned) of baldcypress and water tupelo trees across three forest conditions in 2002 and 2003 (a priori comparison of high vs. sparse using Fisher’s LSD, a = 0.05, for each bar, clipped n = 84 and n = 24 for litter, 1 S.E.).
Increased basal area growth of both tree species with fertilizer application indicates that moderately stressed areas surrounding Lake Maurepas should benefit most from nutrient enhancement associated with a river diversion. Even severely declining swamp, growth of baldcypress trees should increase. However, water tupelo trees, many of which have severely degraded canopies, are not likely to grow in response to nutrient influxes. In the slight decline class, competition with neighboring tree species may limit basal area growth, masking the effects of
302
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304
Fig. 8. Nitrogen content of clipped leaves (left) and leaf litter (right) in 2002 and 2003 correlated with individual defoliation of baldcypress trees (95% CI).
nutrient augmentation. Increased nutrients may have benefited trees in other ways not detectable by aboveground productivity, for example, through increased nutrient storage, photosynthetic rates, belowground biomass, etc. (Webb, 1978). Although traditional studies often address defoliators as pests, recent studies indicate that arthropods affect forest ecosystems in complex ways. Trees are capable of compensatory growth following defoliator outbreaks, often more than replacing growth loss during defoliation (Alfaro and MacDonald, 1988; Alfaro and Shepherd, 1991; Lovett and Tobiessen, 1993; Trumble et al., 1993). Schowalter et al. (1991) found that defoliators removing 20% of the foliage from young Douglasfir doubled the amount of precipitation reaching the forest floor and increased N, K, and Ca flow to the forest floor by 20–30%. Schowalter et al. (1991) reported that three taxa of litter arthropods were more abundant under defoliated trees, suggesting that defoliation could also increase species diversity and stimulate litter decomposition. Similarly, Progar et al. (2000), Schowalter (1992), and Schowalter et al. (1992, 1998)
demonstrated that saprophytic insects and fungi, including bark beetles and some pathogenic fungi, contribute significantly to nutrient cycling from decomposing wood, and enhance the fertility of underlying soil. Changes in detrital processing and arthropod abundance as a result of defoliation in other forested systems implies that defoliation may be important in forested wetlands in Louisiana affecting, detrital processing and aquatic food webs. The results of this study in Lake Maurepas Swamps ascertained that leaf litter biomass was lowered by defoliation, but defoliation increased overall biomass of litterfall reaching the forest floor when insect frass from the spring was included, particularly for baldcypress. Increased litter reaching the forest floor as a result of defoliation could be particularly important in nutrient limited systems, like Maurepas (Lane et al., 2003). Furthermore, these results indicate that individual baldcypress trees contribute more to the detrital production than water tupelo trees, suggesting that baldcypress may be more influential in nutrient cycling processes.
Fig. 9. Nitrogen content of frass (white), clipped leaves (patterned), and leaf litter (black) for baldcypress and water tupelo trees in 2002 only (Fisher’s LSD, a = 0.05, 1 S.E.).
Fig. 10. Nitrogen content of frass (white), clipped leaves (patterned), and leaf litter (black) for both baldcypress and water tupelo trees at the three decline classes in 2002 only (Fisher’s LSD, a = 0.05, 1 S.E.).
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304 Table 3 Overall mean (1 S.E.) percent nitrogen in frass, clipped leaves, and leaf litter (fallen leaves only) associated with baldcypress and water tupelo trees from 2001 to 2003 in the Lake Maurepas basin Variables
Means
(1 S.E.)
Nitrogen (%) Frass (2001–2003) Clipped leaves (2002–2003) Leaf litter (2002–2003)
Baldcypress 1.171 0.026 1.524 0.032 1.353 0.042
Water tupelo 1.757 0.033 1.632 0.033 1.120 0.042
Fertilizer applications increased foliar nitrogen levels across all forest decline classes, both tree species, and leaf types suggesting that the Lake Maurepas Swamp is nitrogen limited in a wide array of forest decline classes. Both predominant tree species therefore have the capacity to store more organic nitrogen in the form of litter if these swamps are nutrient enhanced (via a diversion), with the N loading rate between of 229–458 m3/s diversion rate. Further, a positive correlation between nitrogen levels in clipped leaves and forest health further supports the hypothesis that nitrogen is a limiting factor in this system. Nitrogen was higher in baldcypress leaf litter than water tupelo leaf litter, suggesting that baldcypress might be a better source of organic nutrients for aquatic food webs and other nutrient-cycling ecosystem processes, depending on overall biomass and decomposition rates. Nitrogen content in water tupelo frass was higher than in baldcypress frass, which could result in this being an important source of nitrogen input during the spring. However, inputs from both frass and leaf litter must be considered in context with total biomass contributions. If BCLR produces a larger quantity of frass than FTC, as suggested from the results, total overall nitrogen loading could be higher during years of peak defoliation. Interestingly, defoliation was positively correlated to nitrogen content in leaf litter, suggesting that defoliators not only increase organic nitrogen loading rates in the spring, but could be causing changes in foliar chemistry that direct a higher nitrogen storage in the leaf litter, as opposed to utilizing resources for other storage organs and biochemical pathways (reproductive structures, roots, wood, defensive compounds, photosynthetic rates, etc.). Average nitrogen levels in clipped leaves for baldcypress in this study were 1.52% 0.032, which is comparable to baldcypress foliage in other swamps at 1.43% (Schlesinger, 1978) and 1.52–1.71% (Meeker, 1992). The increased nutrient levels found in the canopy foliage following fertilization suggest that defoliator populations could potentially be increased. By elucidating patterns of nutrient uptake into leaf tissue and frass (by products), this study has provided results explaining nutrient dynamics not previously addressed for forested wetlands. Furthermore, this information can be applied to the ecosystem as a whole to evaluate the potential of various nutrient loading rates based on forest decline, tree species composition, and defoliation rates (e.g. system models). Scientists studying wetland restoration, especially in the Lake Maurepas and Pontchartrain Basins, will be able to use this information for water quality, fisheries, migratory birds, and
303
other wildlife work in coastal Louisiana and in other wetland ecosystems, and will be able to more accurately predict and assess forested wetland processes affected by variation in nutrients and, often, concurrent insect herbivory. Acknowledgements We would like to thank U.S. Forest Service, Region-8 Forest Health, for providing funding to accomplish this research and LSU Ag Chemistry for conducting nutrient analyses. Others that helped were Ankit Modi, Jay Curole, Wood Johnson, Tessa Bauman, Brendan Watson, Anil Surathu, Brandi Johnson, Chandra Dee Glass (all LSU), and Gary Shaffer (Southeastern Louisiana University). We thank Jay Geaghan (LSU) for assistance in statistical analyses; and Seth Johnson (LSU), Dorothy Prowell (LSU), Michael Stout (LSU), James R. Meeker (U.S. Forest Service), and an anonymous reviewer for their thoughtful comments on earlier versions of this manuscript. This manuscript was approved by the Director of the Louisiana Agricultural Experiment Station as manuscript number 06-26-0133. References Abrahamson, L.P., Harper J.D., 1973. Microbial insecticides control forest tent caterpillar in southwestern Alabama. For. Serv. Res. Note (US) SO-157, p. 3. AOAC, 1995. Official Methods of Analysis of the AOAC, 16th ed. Alfaro, R.I., MacDonald, R.N., 1988. Effects of defoliation by the western false hemloc looper on Douglas-fir tree-ring chronologies. Tree-Ring Bull. 48, 3– 11. Alfaro, R.I., Shepherd, R.F., 1991. Tree-ring growth of interior Douglas-fir after one year’s defoliation by Douglas-fir tussock moth. For. Sci. 37, 959–964. Allen, J.A., Chambers, J.L., Pezeshki, S.R., 1997. Effects of salinity on baldcypress seedlings: physiological responses and their relation to salinity tolerance. Wetlands 17 (2), 310–320. Baumann, R.J., Day, J.W., Miller, C.A., 1984. Mississippi deltaic wetland survival: sedimentation versus coastal survival. Science 224, 1093–1095. Boesch, D.F., Josselyn, M.N., Mehta, A.J., Morris, J.T., Nuttle, W.K., Simenstad, C.A., Swift, D.J.P., 1994. Scientific assessment of coastal wetland loss, restoration and management in Louisiana. J. Coastal Res. (Special issue no. 20), 20. Bormann, F.H., Likens, G.E., 1979. Pattern and Process in a Forested Ecosystem. Springer-Verlag, NewYork, NY, p. 253. Braun, D.M., Goyer, R.A., Lenhard, G.J., 1989. Biology and mortality agents of the fruit tree leafroller (Lepidoptera:Tortricidae), on baldcypress in Louisiana. J. Entomol. Sci 25, 176–184. Brown, C.A., Montz, G.N., 1986. Baldcypress and the Tree Unique, the Wood Eternal. Claitor’s Publishing Division, Baton Rouge, LA, USA, p. 139. Christenson, L.M., Lovett, G.M., Mitchell, M.J., Groffman, P.M., 2002. The fate of nitrogen in gypsy moth frass deposited to an oak forest floor. Oecologia 131 (3), 444–452. Conner, W.H., Brody, M., 1989. Rising water levels and the future of southeastern Louisiana swamps. Estuaries 12, 318–323. Conner, W.H., Buford, M.A., 1998. Southern deepwater swamps. In: Messina, M.G., Conner, W.H. (Eds.), Southern Forested Wetlands: Ecology and Management. Lewis Publishers, NewYork, NY, pp. 261–287. Conner, W.H., Day, J.W., Jr., 1984. The impacts of increased flooding on commercial wetland forests in the Lake Verret watershed. Final Report of Project 83-LBR/022-B to Louisiana Board of Regents. Baton Rouge, LA, USA. Conner, W.H., Day Jr., J.W., 1988. Rising water levels in coastal Louisiana: implications for two coastal forested wetland areas. J. Coastal Res. 4, 589– 596.
304
R.S. Effler et al. / Forest Ecology and Management 236 (2006) 295–304
Conner, W.H., Toliver, J.R., 1990. Long-term trends in the baldcypress (Taxodium distichum) resource in Louisiana (USA). For. Ecol. Manage. 33–34, 543–557. Coulson, R.N., Witter, J.A., 1984. Forest Entomology: Ecology and Management. Wiley Publishing Co., NY, p. 699. Day Jr., J.W., Shaffer, G.P., Britsch, L.D., Reed, D.J., Hawes, S.R., Cahoon, D., 2000. Pattern and process of land loss in the Louisiana coastal zone: an analysis of spatial and temporal patterns of wetland habitat change. Estuaries 23, 425–438. Day Jr., J.W., Arancibia, A.Y., Mitsch, W.J., Lara-Dominguez, A.L., Day, J.N., Ko, J.Y., Lane, R., Linsey, J., Lomeli, D.A., 2003. Using ecotechnology to address water quality in wetland habitat loss problems in the Mississippi basin: a hierarchical approach. Biotechnol. Adv. 22, 135–159. Day Jr., J.W., Conner, W.H., Shaffer, G.P., 2006. The importance of pulsed physical events for sustainability of Louisiana coastal forested wetlands. In: Nettles, J., Williams, T. (Eds.), Hydrol. Manage. Forested Wetlands Proc. Int. Conf., American Society of Agricultural and Biological Engineers, St. Joseph, MI, pp. 156–158. Dennis, J.V., 1988. The Great Cypress Swamps. Louisiana State University Press, Baton Rouge, LA, USA, p. 142. Goyer, R.A., Chambers, J., 1996. Evolution of insect defoliation in baldcypress and its relationship to flooding. USDI. NBS. Biol. Sci. Rep. 8. Goyer, R.A., Lenhard, G.J., 1988. A new insect pest threatens baldcypress. LA Agriculture, LA Agricultural Experiment Station 31, pp. 16–17, 21. Hutchens, J.J., Benfield, E.F., 2000. Effects of forest defoliation by the gypsy moth on detritus processing in southern Appalachian streams. Am. Midland Nat. 143 (2), 397–404. Johnson, C.W., 2004. The performance of the baldcypress leafroller (Archips goyerana Kruse, Lepidoptera:Tortricidae) in response to fertilization, thinning, and genetic variation in host baldcypress (Taxodium distichum L. Richard). MS Thesis. Louisiana State University, Baton Rouge, LA. Kemp, P.G., Conner, W.H., Day Jr., J.W., 1985. Effects of flooding on decomposition and nutrient cycling in a Louisiana swamp forest. Wetlands 5, 35– 50. Klock, G.O., Wickman, B.E., 1978. Ecosystems effects. In: Brookes, M.H., Stark, R.W., Campbell, R.W. (Eds.), The Douglas-Fir Tussock Moth: A Synthesis. USDA For. Ser. Technical Bull. 1585, pp. 90–94. Lane, R.R., Day, J.W., Thibodeaux, B., 1999. Water quality analysis of a freshwater diversion at Caernarvon, Louisiana. Estuaries 22, 327– 336. Lane, R.R., Mashriqui, H.S., Kemp, G.P., Day, J.W., Day, J.N., Hamilton, A., 2003. Potential nitrate removal from a river diversion into a Mississippi delta forested wetland. Ecol. Eng. 20, 237–249. Louisiana Coastal Wetlands Conservation and Restoration Task Force, 1993. Louisiana Coastal Wetlands Plan. Louisiana Wetlands Coastal Planning. Protection, and Restoration Act, Baton Rouge, LA, USA. Louisiana Coastal Wetlands Conservation and Restoration Task Force and the Wetlands Conservation and Restoration Authority, 1998. Coast 2050: Toward a Sustainable Louisiana. Louisiana Department of Natural Recources, Baton Rouge, LA, USA. Lovett, G., Tobiessen, P., 1993. Carbon and nitrogen assimilation in red oaks (Quercus rubra L.) subject to defoliation and nitrogen stress. Tree Physiol. 12, 259–269. Mancil, E., 1980. Pullboat logging. J. For. Hist. 24, 135–141. Meeker, J.R., 1992. Host quality of baldcypress and its influence of the fruit tree leafroller, Archips agyrospila (Walker) (Lepidoptera: Tortricidae), Performance in Forested Wetlands of Louisiana. A Dissertation. Louisiana State University. Baton Rouge, LA, USA. Mitsch, W.J., Gosselink, J.G., 1993. Wetlands, second ed. Van Norstrand Reinhold, New York, NY. Mossa, J., 1996. Sediment dynamics in the lowermost Mississippi River. Eng. Geol. 45, 457–479. Nachod, L.H., 1977. Spring defoliation by forest insects in Louisiana. Insect and Disease Report. Louisiana Office of Forestry, Woodworth, LA, USA, p. 2. Nachod, L.H., Kucera, D.R., 1971. Observations of the forest tent caterpillar in south Louisiana. Insect and Disease Report. Louisiana Office of Forestry, Woodworth, LA, USA, p. 2.
Progar, R.A., Schowalter, T.D., Morrell, J.J., Freitag, C.M., 2000. Respiration from coarse woody debris as affected by moisture and saprotroph functional diversity in western Oregon. Oecologia 124, 426–431. Salinas, L.M., DeLaune, R.D., Patrick Jr., W.H., 1986. Changes occurring along a rapidly emerging coastal area: Louisiana, USA. J. Coastal Res. 2, 269– 284. SAS Institute, Inc., 1999. SAS/STAT User’s Guide Version 8.0. Cary, NC. Schlesinger, W.H., 1978. Community structure, dynamics and nutrient cycling in the Okefenokee cypress swamp-forest. Ecol. Monogr. 48, 43–65. Schowalter, T.D., 1981. Insect herbivore relationship to the state of the host plant: biotic relation of ecosystem nutrient cycling through ecosystem succession. Oikos 37, 126–130. Schowalter, T.D., 1992. Early decomposition and nutrient dynamics of oak (Quercus) logs at four sites across a North American gradient. Can. J. For. Res. 22, 161–166. Schowalter, T.D., Caldwell, B.A., Carpenter, S.E., Griffiths, R.P., Harmon, M.E., Ingham, E.R., Kelsey, R.G., Lattin, J.D., Moldenke, A.R., 1992. Decomposition of fallen trees: effects of initial conditions and heterotroph colonization rate. In: Singh, K.P. (Ed.), Tropical Ecosystems: Ecology and Management. Wiley Eastern, Ltd., New Delhi, pp. 371–381. Schowalter, T.D., Crossley Jr., D.A., 1983. Forest canopy arthropods as sodium, potassium, magnesium and calcium pools in forests. For. Ecol. Manage. 7, 143–148. Schowalter, T.D., Sabin, T.E., Stafford, S.G., Sexton, J.M., 1991. Phytophage effects primary production, nutrient turnover, and litter decomposition on young Douglas-fir in western Oregon. For. Ecol. Manage. 42, 229–243. Schowalter, T.D., Zhang, Y.L., Sabin, T.E., 1998. Decomposition and nutrient dynamics of oak (Quercus spp.) logs after five years of decomposition. Ecography 21, 3–10. Seastedt, T.R., Crossley Jr., D.A., 1984. The influence of arthropods on ecosystems. BioScience 34, 157–161. Shaffer, G.P., Perkins, T.E., Hoeppner, S.S., Howell, S., Beneard, T.H., Parsons, A.C., 2003. Ecosystem health of the Maurepas Swamp: feasibility and projected benefits of a freshwater diversion. Final Report. Environmental Protection Agency, Dallas, TX, Region 6, p. 105. Souther-Effler, R.F., 2004. Interactions of insect herbivory and multiple abiotic stress agents on two wetland tree species in southeast Louisiana swamps. PhD Dissertation. Louisiana State University, Baton Rouge, LA, USA. Swank, W.T., Waide, J.B., Crossley Jr., D.A., Todd, R.L., 1981. Insect defoliation enhances nitrate export from forest ecosystems. Oecologia 51, 297– 299. Templet, P.H., Meyer-Arendt, K.J., 1988. Louisiana wetland loss: regional water management approach to the problem. Environ. Manage. 12, 181– 192. Trumble, J.T., Kolodny-Hirsch, D.M., Ting, I.P., 1993. Plant compensation for arthropod herbivory. Annu. Rev. Entomol. 38, 93–119. Turner, R.E., 1997. Wetland loss in the northern Gulf of Mexico: multiple working hypotheses. Estuaries 20, 1–13. Thompson, D.A., 2000. The influence of Hydrological alterations upon wetland hydrodynamics and plant growth on the Manchac Landbridge, Southeastern Louisiana, USA. Master’s Southeastern Louisiana University, Hammond, LA. Visser, J.M., 1989. The impact of vertebrate herbivores on primary production of Sagittaria marshes in the Wax Lake Delta, Atchafalaya Bay, Louisiana. PhD Dissertation. Louisiana State University, Baton Rouge, LA, USA. Webb, W.L., 1978. Effects of defoliation and tree energetics. In: Brookes, M.H., Stark, R.W., Campbell, R.W. (Eds.), The Douglas-Fir Tussock Moth: A Synthesis. USDA For. Serv. Technical Bull. 1585, pp. 77–81. White, W.B., Schneeberger, N.F., 1981. Socioeconomic impacts. In: Doane, C.C., McManus, M.L. (Eds.). The Gypsy Moth: Research Toward Integrated Pest Management. USDA For. Ser. Technical Bull. 1584, pp. 681–694. Wilson, L., and Associates, Shaffer, G., Hester, M., Kemp, P., Mashriqui, H., Day, J., Lane, R., 2001. Diversion into Maurepas Swamp. Project Report to U.S. Environmental Protection Agency, Region 6, Dallas, TX Contract No. 68-06-0067, WA #5-02, June 2001. Zlotin, R.I., Khodashova, K.S., 1980. In: French, N.R. (Ed.), The Role of Animals in Biological Cycling of Forest-Steppe Ecosystems (English translation). Dowden, Hutchinson, and Ross, Stroudsburg, PA, p. 221.