Next Article in Journal
Stress Wave Hybrid Imaging for Detecting Wood Internal Defects under Sparse Signals
Next Article in Special Issue
Analysis of the Environments Where Natural Regeneration Is Established in the Absence of a Wildfire in the Open Pinus brutia Forests in the Middle Elevations of the Central Part of Cyprus
Previous Article in Journal
SIMCB-Yolo: An Efficient Multi-Scale Network for Detecting Forest Fire Smoke
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Forests
Volume 15
Issue 7
10.3390/f15071138
forests-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Pinus banksiana Lamb. Regeneration Patterns in a Lacustrine Dune System

by
Jonathan C. Danielson
1,2,
Adam R. Warrix
1,3,
Madison E. Lehman
1,
Andrew C. Lehman
1 and
Jordan M. Marshall
1,*
1
Department of Biological Sciences, Purdue University Fort Wayne, Fort Wayne, IN 46805, USA
2
Indiana Institute of Technology, Fort Wayne, IN 46803, USA
3
R. Nelson Snider High School, Fort Wayne, IN 46815, USA
*
Author to whom correspondence should be addressed.
Forests 2024, 15(7), 1138; https://doi.org/10.3390/f15071138
Submission received: 31 May 2024 / Revised: 19 June 2024 / Accepted: 27 June 2024 / Published: 29 June 2024
(This article belongs to the Special Issue Impact of Disturbance on Forest Regeneration and Recruitment)
Browse Figures
Versions Notes

Abstract

:
Successional patterns in lacustrine sand dunes along Lake Superior begin with grass-dominated plant communities leading to the establishment of Pinus banksiana Lamb. as initial forests. Using maximum entropy models, we predicted P. banksiana seedling and sapling patterns within the Grand Sable Dunes, Pictured Rocks National Lakeshore, USA, based on slope, aspect, forest basal area, and vegetation types. Across the different vegetation types, there were variable probabilities of seedling and sapling occurrence. For both seedlings and saplings, the higher likelihoods of occurrence were observed in coastal pine barrens vegetation types. P. banksiana regeneration is occurring in the Grand Sable Dunes in the absence of fire, with seedlings establishing and saplings being recruited in a variety of vegetation types. With the greatest probabilities in barrens, there is likely a relationship with seed source and canopy density. Understanding regeneration patterns in dune ecosystems is necessary to predict the future forest arrangement and colonization of P. banksiana into the dunes. These results contribute insights into the dynamics of plant communities in lacustrine dune systems, specifically the establishment of P. banksiana seedlings in various vegetation types. Continued forest establishment and increasing P. banksiana density will influence endangered species and non-native species management strategies for Pictured Rocks National Lakeshore.

1. Introduction

Lacustrine sand dune ecosystems along the Great Lakes in North America receive repeated natural disturbance during both the growing and non-growing seasons [1]. Plant communities within these dune systems are categorized as globally and state vulnerable due to human threats from recreation, development, and invasive species [2,3,4]. Even with regular and recurring sand accretion and excavation due to disturbances, dune surface stabilization occurs through successional processes via multiple plant species colonization [5]. Cowles [6] provided initial descriptions of succession and the changing plant communities in the Great Lakes basin along Lake Michigan, drawing from previous work by Warming [7]. Since those early descriptions, plant communities within sand dunes have often been described as a chronosequence, from bare beach to closed canopy forest, which represents both the distance from the shore and the time since disturbance [5]. However, there has been criticism of the chronosequence concept as a simplification of dune ecosystem reality [8].
Following open-dune grass communities, dune stabilization is marked by increases in the establishment of woody species, including shrubs and trees [9]. Establishment and recruitment of woody species is likely driven by combinations of soil development, herbaceous plant stabilization, and stochastic seed dispersal [10,11,12]. Burial by sand creates limitations for woody species establishment, forming zones of colonization where active sand movement limits early-successional establishment; stabilization by mid- and overstory woody plants, however, increases in later-successional areas [13].
In sand dune ecosystems along the Great Lakes shorelines, Pinus banksiana Lamb. (jack pine) is a common tree species that initiates forests as an earlier-successional stage species [14,15]. P. banksiana often responds positively to disturbances; however, the successional pathway following that response is variable [16]. Young P. banksiana stands have important ecological value, providing habitat for specific bird communities, unlike older P. banksiana stands [17]. Additionally, as P. banksiana stands age, disturbance mechanisms change. P. banksiana is a fire-adapted species, exhibiting traits that increase regeneration after fire [18]. Mature P. banksiana release the majority of seeds within a few growing seasons post-fire leading to immediate regenerative response [19]. However, variability in fire severity leads to variable patterns in P. banksiana success [20]. Additionally, frequency of cone serotiny has phenotypic variation within and between populations related to fire disturbance [21]. There is phenotypic plasticity in P. banksiana cone serotiny, as the trait is heritable and there is a range of serotiny in individuals [22,23]. When fire is a common disturbance within P. banksiana forests, there is an increase in the frequency of high serotinous cones [21,24]. Moreover, when trees are at higher density (i.e., increased probability that crown fires will pass between individuals), there is also an increase in high serotiny frequency [24]. Conversely, when other disturbances are more common and when trees are spaced farther apart, low serotinous cones occur more frequently [21]. The lack of serotiny, especially in smaller P. banksiana individuals, may be important to increase regeneration in systems lacking fire or with long fire return intervals [25].
Predicting natural regeneration patterns (i.e., seedling and sapling dispersion) provides insight into the potential future forest structure and composition following successional or other stand dynamics pathways [26]. Such predictions can include responses to disturbances, as well as habitat suitability alterations related to climate change [26,27,28]. Species distribution models are a tool for making predictions about probabilities of occurrence using spatial data. While nearly any valid mathematical modeling approach can be used from general linear to ordination techniques to produce probability maps [29], maximum entropy has become a popular approach [30]. Maximum entropy modeling is a technique to make predictions based on more complicated parameters, more complicated interactions among parameters, and incomplete data, potentially better than other modeling techniques [31,32]. Such models can incorporate both continuous and categorical data into a model, which will approach the maximum entropy probability distribution [32,33]. The expectation of selecting the model with the maximum entropy is that it will be the probability distribution closest to reality with the least bias, which is due to the entropy (i.e., uncertainty) being uniformly distributed [34,35]. Because of the importance of P. banksiana regeneration in the initiation of forests within lacustrine dune systems in North America, we used maximum entropy to predict seedling and sapling patterns.
The objectives of this study were to (1) map locations of P. banksiana seedlings within Grand Sable Dunes, Pictured Rocks National Lakeshore, USA, (2) predict regeneration patterns using maximum entropy modelling techniques with environmental data layer inputs of slope, aspect, forest canopy height, forest basal area, and vegetation types, and (3) test the hypothesis that earlier successional portions of the dunes will have higher probability of P. banksiana occurrence.

2. Materials and Methods

2.1. Study Site

The Grand Sable Dunes is an approximately 900 ha perched dune field on the shore of Lake Superior, in Michigan, USA, and is managed as part of Pictured Rocks National Lakeshore (46°39′22″ N, 86°02′58″ W; Figure 1A). Within this dune ecosystem, areas exist of active disturbance and sand movement, interdunal areas with various stages of stabilization, and back dune areas with well-established coniferous and hardwood forest stands [36,37]. Past fires have occurred in the Grand Sable Dunes; however, those were likely scattered and were not repeated at a given site [38]. Several plant species categorized as state or federally endangered, threatened, or of concern, occur within the Grand Sable Dunes, and are likely influenced by dune stabilization and successional processes [9,37].

2.2. Model Development

In order to collect model development data across vegetation cover types (e.g., forest, herbaceous, open dune areas), we employed three methods to locate plots within the Grand Sable Dunes. In June 2018, we established three transects perpendicular to the shore that extended 1500 m from access points along the southern boundary road (Figure 1B). Each transect was a 2 m wide belt and root collar diameter was measured for all P. banksiana individuals (≤15.0 mm in diameter). Additionally, the location along the transect for each individual P. banksiana was also measured. In July 2020 and 2021, we randomly selected plot locations on a 50 m spaced grid overlaid on the Grand Sable Dunes (n = 50 and 174, respectively; Figure 1C,D). The 2021 survey was focused on the eastern half of the Grand Sable Dunes, which has the greatest abundance of P. banksiana. In July 2022, also on a 50 m grid, we selected points along transects that appeared to traverse from open dune to forest (n = 118; Figure 1E). For surveys in 2020–2022, we measured all species, including P. banksiana, seedling (≤15.0 mm root collar diameter) within a 2.5 m radius circular plot.
To characterize the forest stands, we measured the basal area by species using a 10-factor prism. In 2018, this was carried out at sampling points on 50 m intervals along each transect to reduce the likelihood of overlapping trees within survey locations. From 2020 to 2022, basal area was measured from the center of each plot location. Throughout all years, when P. banksiana was encountered in the basal area counts, vigor was categorized similar to Millers et al. [39] (1–6 rating scale; 1 was a healthy tree with relatively few dead twigs, foliage density and normal color; 2 had occasional large dead branches in the upper portion, foliar density below normal; 3 described a crown that demonstrated moderate dieback, several large branches in upper crown, bare twigs showing; 4 represented approximately half of the crown being dead; 5 represented over half of the crown being dead; and 6 was a standing dead tree) and Choristoneura pinus pinus Freeman (jack pine budworm) damage was noted as presence/absence (1/0).
We defined P. banksiana seedlings as having a root collar diameter of ≤5 mm and saplings as having a root collar diameter of >5 mm. For both seedlings and saplings, we divided the locations with P. banksiana individuals present into two similarly sized groups: one for model development and the other for model testing. Independent variables for model development were slope and aspect, vegetation cover type, forest canopy height, P. banksiana basal area, and all other species basal area. Slope and aspect were derived from a digital elevation model with 1 m2 pixel size [40]. Vegetation cover types were rasterized from polygon vector data [41]. Forest canopy height was derived from light detection and ranging (lidar) point cloud data [42] and was normalized using a TIN interpolation of ground categorized data points and canopy height was estimated using a pitfree technique, as described by Khosravipour et al. [43], with functions in the lidR package in R [44,45,46]. P. banksiana basal area raster was produced using basal area measurements from plot locations pooled across years and ordinary co-kriging (a Matern model with forest canopy height as a co-variable) with the gstat package in R [47,48]. Similarly, a raster of basal area for other species was produced using the basal area of all species other than P. banksiana and ordinary co-kriging (a Matern model with forest canopy height as a co-variable). All independent variable rasters were resampled based on the aspect raster (i.e., 1 m2 pixel).
Prior to developing models, we tested collinearity of independent variables using the ‘pairs’ function in the terra package in R [49]. A decision was made to omit forest canopy height form the model development based on correlation with vegetation type and its use in co-kriging to produce the basal area rasters (Figure S1). We used maximum entropy models to predict probabilities of P. banksiana seedling and sapling occurrences from the model development data with aspect, other species basal area, P. banksiana basal area, slope, and vegetation cover type (Figure S2). Maximum entropy models were developed using the ‘MaxEnt’ function in the predicts package as a wrapper for Maxent software (version 3.4.3) [50] with jackknifing in R [44]. Vegetation cover was included as a ‘factor’ option as the data were categorical cover types. Additionally, the ‘MaxEnt’ function was run with the ‘removeDuplicates=T’ option to remove duplicate occurrences within the same independent variable pixel, which was a possibility along the 2018 transects (i.e., occurrences within the same 1 m2 pixel), and background points were equal to the 2 * number model development occurrences for each of the seedling and sapling models. Default values for all other options within the ’MaxEnt’ function were used.

2.3. Model Testing

We tested model success using both P. banksiana seedling and sapling occurrence (model testing data as presence) and absence data (plots where zero P. banksiana were encountered in the period of 2020–2022 as absence). With these presence and absence data, we used the ‘pa_evaluate’ and ‘threshold’ functions in the predicts package to identify maximum kappa values as our probability threshold to test model success [50,51]. We extracted the predicted probability of P. banksiana occurrence at the model testing locations and coded predicted occurrence based on the threshold kappa (i.e., if the probability was >kappa, it was coded as predicted presence; if the probability was ≤kappa, it was coded as predicted absence). This provided us with true positive (observed presence and predicted presence), true negative (observed absence and predicted absence), false positive (observed absence but predicted presence), and false negative (observed presence but predicted absence). From these true and false positive and negative counts, we used a McNemar test for the hypothesis that the false negative and false positive values were equal. We also calculated accuracy to aid in characterizing the overall model success.

2.4. Statistical Analysis

All model development and testing analyses were conducted in R version 4.3.2 [44]. Evaluation of the models using the ‘pa_evaluate’ function in the predicts package provided the area under the receiver operating characteristic curve (AUC) as a gauge of model accuracy with the model testing data. We used partial dependency plots to visualize the marginal effect of each independent variable in the maximum entropy models for seedling and saplings using ‘partialResponse’ function in the predicts package.

3. Results

In 2018, along the three 1.5 km transects, we measured 1251 P. banksiana individuals. We measured a total of 120 seedlings and saplings in 2020, 3361 in 2021, and 1964 in 2022 (Table 1, Figure 2). For P. banksiana, seedlings (≤5 mm root collar diameter) accounted for 85.1% of individuals of that species in 2018, 70.7% in 2020, 55.7% in 2021, and 50.8% in 2022. Overall, P. banksiana accounted for 11.3% of seedlings and 15.4% of saplings for 2020–2022 for all species measured (Figure 3). We observed signs and symptoms of C. pinus pinus infestation across years, with a mean occurrence of 12.3% (±SD 9.2%) of mature P. banksiana with infestation. The year with the greatest C. pinus pinus occurrence was 2021 (22.3% of trees) and the year with lowest was 2022 (4.2%). In all cases, we observed signs of C. pinus pinus presence on each tree as a single or relatively few branch tips; extensive defoliation of individual trees was not observed.

3.1. Model Development

The seedling model was produced using 565 presence points and 1130 background points (Equation (S1)). Other species basal area had the greatest permutation importance on the seedling model, followed by vegetation type (Table 2). The visualized marginal effect for other species basal area showed a seedling peak at zero basal area, with a steep decrease in probability as basal area increased (Figure 4B). The peak seedling probability for vegetation type corresponded to the categorical code for ’Great Lakes Coast Pine Barrens (barrens phase)’ (Figure 4D).
The sapling model was produced using 135 presence points and 270 background points (Equation (S2)). P. banksiana basal area had the greatest permutation importance followed by other species basal area (Table 2). Marginal effect for P. banksiana basal area peaked at approximately 30–35 m2/ha (Figure 4F). The marginal effect of other species basal area followed a pattern similar to the seedling model, with peak sapling probability observed at zero basal area. Similarly, vegetation type peaked at the ’Great Lakes Coast Pine Barrens (barrens phase)’ category (Figure 4H).
For the seedling model, AUC was 0.792 with a maximum kappa of 0.61 and for the sapling model, AUC was 0.757 and 0.58 maximum kappa. Predicted seedling and sapling probabilities using the independent variables resulted in the lowest values on the southern boundary and greater values more central to the dunes (Figure 5).

3.2. Model Testing

Using the maximum kappa as our predicted presence threshold, we tested the seedling and sapling models. Of the 342 plots surveyed in the period of 2020–2022, P. banksiana seedlings were present at only 66 plots. This provided us with 276 absence locations, in addition to 565 presence locations from 2018 and 2020–2022. The tested seedling model resulted in an AUC of 0.779 (similar to the model development AUC), suggesting that the model was able to distinguish true and false locations between P. banksiana seedlings 78% of the time. Using the 0.61 maximum kappa defined in the model evaluation as a threshold, the model testing data were defined as predicted presence and absence (1, 0) and a McNemar’s chi-square test was used to test for the incorrect classification expected presence and absence. The hypothesis was rejected ( X 1 2 = 74.28, p < 0.001), suggesting that observed and predicted presences and absences were misaligned, with 13.1% of the true presence locations classified as absence but 46.7% of true absence locations classified as presences. Overall accuracy for the seedling model was 70.3%.
P. banksiana saplings were absent from 256 plots in the period of 2020–2022. We used these absence locations and 136 presence locations (2018 and 2020–2022) to test the model. The sapling model had an AUC of 0.666, which was less than the model development AUC and suggested that the model was only able to distinguish approximately 67% of the true and false locations. Using the 0.58 maximum kappa defined in the model evaluation as a threshold, the McNemar’s chi-square test failed to reject the null hypothesis, suggesting that the observed and predicted presences and absences were not misaligned ( X 1 2 = 0.008, p = 0.930), with 47.8% of presence locations classified as absence and 25.2% of absence locations classified as presence. Overall accuracy for the sapling model was 66.8%.
Mean probabilities for seedling occurrences were greatest in the ‘Great Lakes Coast Pine Barrens’ vegetation types, which included ‘barrens phase’ and ‘woodland phase’ (Figure 6). These ‘Barrens’ types appeared similar to the ‘Jack Pine/Balsam Fir Forest’ type. However, for the sapling occurrences, the greatest mean probability occurred in the ‘Sand Cherry Dune Shrubland’, which was greater than the ’Barrens’ and ’Jack Pine/Balsam Fir Forest’ types (Figure 7).

4. Discussion

Understanding regeneration patterns within a lacustrine dune system provides insight into the future forest that will colonize the dunes. This study aimed to map P. banksiana seedlings and saplings within the Grand Sable Dunes to predict regeneration patterns and identify areas where seedlings are most likely to occur. While P. banksiana is well documented as a fire-adapted species, fire is uncommon within the Grand Sable Dunes. Soil profiles within the dunes provide evidence of relatively rare, scattered fire [38] and the authors of this paper have not observed evidence of fire within the Grand Sable Dunes.
Sand movement is significant in shaping the dune surface and subsequently influencing plant community colonization. With fire lacking in the Grand Sable dunes, disturbances are limited to sand movement. The ‘Great Lakes Beachgrass Dune’ vegetation type would be the earliest successional stage dominated by Ammophila breviligulata, which is common in the region on active foredunes [41,52,53]. Our original hypothesis was that we would observe increases in probabilities of P. banksiana seedling and sapling occurrences in the earlier successional habitats (i.e., beachgrass dune vegetation type). While there were P. banksiana seedlings occurring in this vegetation type, it was not significantly different from other vegetation types. Only probabilities in the two ‘Great Lakes Coast Pine Barrens’ vegetation types appeared greater than the majority of other vegetation types. The barrens phase is 10–25% tree canopy cover, while the woodlands phase is 25–60% tree canopy cover [41]. P. banksiana is considered an early forest in the Grand Sable Dunes, but seedling occurrence was lower in the earliest successional stage.
Seed source is important in determining the variable patterns in P. banksiana seedling and sapling occurrences. Areas with high-density P. banksiana are often immediately adjacent to areas with low-density [54]. Again, we observed high probability of P. banksiana seedling and sapling in the coastal pine barrens vegetation types. With the absence of fire, we observed variability in the occurrence of P. banksiana seedlings and saplings similar to other natural regeneration studies [54,55]. Because of this lack of fire, we may have seen more P. banksiana seedlings and saplings occurring in vegetation types that naturally experience less fire (e.g., ‘Jack Pine/Balsam Fir Forest’). However, the greatest probability of predicted P. banksiana sapling occurrence was in ‘Sand Cherry Dune Shrubland’. Prunus pumila (sand cherry) has variable affinity for fire as well [56]. While fire is not present in the Grand Sable Dunes, the likely presence of phenotypic variation in both P. banksiana and P. pumila along with the occurrence together in the ‘Sand Cherry Dune Shrubland’ is an interesting correlation, which requires future investigation.
P. banksiana regeneration and survival is linked to variation in habitat characteristics from xeric to mesic conditions [15,57]. Additionally, as a shade-intolerant species, there typically is an inverse relationship between overstory shade and seedling survival. Growth, competitive ability, and carbon assimilation rates are relatively high for P. banksiana in open canopy areas, with significant reductions as the canopy increases [58,59,60]. Because of this, we would have expected greater numbers and probabilities of P. banksiana seedlings in more open dune habitats (e.g., ‘Great Lakes Beachgrass Dune’, ‘Great Lakes Coast Pine Barrens [barrens phase]’ vegetation types). These vegetation types have rapidly drained sandy soils [41]. However, we did observe some high probabilities of P. banksiana occurrence in our present test data in the ‘Jack Pine/Balsam Fir Forest’ vegetation type. These areas are characterized by closed canopy with approximately 80% or more cover, but with rapidly drained sandy soils [41]. This ’Jack Pine/Balsam Fir Forest’ type is a later successional stage with recruitment of Abies balsamea (balsam fir). There should be rapid mortality in P. banksiana seedlings with increased shading [61]. As we did not measure percent canopy cover or light levels, we are unsure if the occurrences of P. banksiana seedlings were in specific gaps.
While C. pinus pinus is possibly the most important defoliator of P. banksiana in the Great Lakes region [62], we did not encounter many occurrences of symptoms in our surveys. Outbreak cycles often occur in 5-, 6-, and 10-year intervals, which means we possibly were surveying trees between peak years [63]. The lack of fire in the Grand Sable Dunes and the presence of other disturbances from wind and storms should lead to improved host quality for C. pinus pinus [64]. However, we did not assess the density of pollen cones in our surveys, which positively influences survival of C. pinus pinus larvae [65]. There is likely a close association between P. banksiana, C. pinus pinus, and fire [54]. The low density of P. banksiana stands in Grand Sable Dunes may negatively influence C. pinus pinus infestations [66]. P. banksiana regeneration is occurring within the Grand Sable Dunes without the presence of fire and relatively low density of C. pinus pinus.
Due to the sensitivity of the soils and associated plant communities in the Grand Sable Dunes, active forest management is not likely to occur within this site. However, other management activities related to controlling non-native species do occur with varying success [67,68]. While this study provides limited implications on forest management decisions within Pictured Rocks National Lakeshore, it does provide insight into future forest possibilities as the Grand Sable Dunes stabilize and P. banksiana stands age and regenerate. There appears to be self-perpetuation of the ‘Great Lakes Coast Pine Barrens’ vegetation types, as well as colonization into the ‘Great Lakes Beachgrass Dune’ vegetation types and sapling recruitment in ‘Sand Cherry Dune Shrubland’ vegetation type. These latter two provide insight into P. banksiana stand expansion into new areas of the dunes. Further research is needed to understand the minutia of environmental conditions surrounding the influence of each vegetation type on the regeneration and recruitment of P. banksiana–canopy cover and light availability especially. From that research, the models presented may be refined to further clarify the mechanisms and patterns leading to P. banksiana colonization and stabilization of the Grand Sable Dunes.

5. Conclusions

As a lacustrine dune system, the Grand Sable Dunes in Pictured Rocks National Lakeshore is undergoing active forest establishment with the continued regeneration of P. banksiana. Understanding regeneration of P. banksiana in this system, which lacks fire, is important as the species is regularly described in relation to fire presence. We predicted seedling and sapling occurrence in several vegetation types with higher probabilities in forest barrens types, as well as in other forest types that naturally would experience less fire. While P. banksiana saplings had relatively high predicted probabilities in forest barrens vegetation types, the highest values were in a shrubland type associated with P. pumila. With the continuation of P. banksiana forest and potential expansion into other vegetation types, the dunes will be important in managing the threatened and endangered species within the dunes. As those species have variable stability and shade-tolerance characteristics, it will be important in the future to quantify potential habitat loss or gains in relation to forest colonization within the dunes. Our models are likely applicable to lacustrine dune systems with P. banksiana, especially in the Great Lakes region of North America. However, in other dune systems, plant colonization in response to disturbances is still important to successional processes. Similar models may be necessary to understand those processes with different plant species.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/f15071138/s1, Figure S1: Pairs plot between aspect, forest canopy height, slope, and vegetation type; Figure S2: Pairs plot between aspect, Pinus banksiana basal area, other species basal area, slope, and vegetation type; Equation (S1): Maximum entropy model equations for probability of Pinus banksiana seedlings; Equation (S2): Maximum entropy model equations for probability of Pinus banksiana saplings.

Author Contributions

Conceptualization, J.M.M.; methodology, J.C.D., A.R.W. and J.M.M.; investigation, J.C.D., A.R.W., M.E.L., A.C.L. and J.M.M.; formal analysis, J.M.M.; writing—original draft preparation, J.M.M.; writing—review and editing, J.C.D., A.R.W., M.E.L., A.C.L. and J.M.M.; supervision, J.M.M. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Original data are available at the Purdue University Research Repository at https://doi.org/10.4231/QP1T-SC36. All other data available from cited sources.

Acknowledgments

Field data collection carried out under National Park Service Scientific Research and Collecting Permits PIRO-2018-SCI-0003, PIRO-2020-SCI-0002, PIRO-2021-SCI-0001, and PIRO-2022-SCI-0001.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Marsh, W.M.; Marsh, B.D. Wind erosion and sand dune formation on high Lake Superior bluffs. Geogr. Ann. Ser. Phys. Geogr. 1987, 69, 379–391. [Google Scholar] [CrossRef]
  2. McLeod, K.W.; Murphy, P.G. Establishment of Ptelea trifoliata on Lake Michigan sand dunes. Am. Midl. Nat. 1977, 97, 350–362. [Google Scholar] [CrossRef]
  3. Kost, M.A.; Albert, D.A.; Cohn, J.G.; Slaughter, B.S.; Schillo, R.K.; Weber, C.R.; Chapman, K.A. Natural Communities of Michigan: Classification and Description; Michigan Natural Features Inventory: Lansing, MI, USA, 2007; Report No. 2007-21. [Google Scholar]
  4. Marshall, J.M.; Storer, A.J.; Leutscher, B. Comparative analysis of plant and ground dwelling arthropod communities in lacustrine dune areas with and without Centaurea biebersteinii (Asteraceae). Am. Midl. Nat. 2008, 159, 261–274. [Google Scholar] [CrossRef]
  5. Lichter, J. Primary succession and forest development on coastal Lake Michigan sand dunes. Ecol. Monogr. 1998, 68, 487–510. [Google Scholar] [CrossRef]
  6. Cowles, H.C. The ecological relations of the vegetation on the sand dunes of Lake Michigan. Bot. Gaz. 1899, 27, 95–117, 167–202, 281–308, 361–391. [Google Scholar] [CrossRef]
  7. Warming, E. Botaniske Exkursioner 2. De psammophile Formationer i Danmark. Vidensk. Meddelel. fra den Naturhist. Foren. 1891, 1891, 153–202. [Google Scholar]
  8. Johnson, E.A.; Miyanishi, K. Testing the assumptions of chronosequences in succession. Ecol. Lett. 2008, 11, 419–431. [Google Scholar] [CrossRef]
  9. Marshall, J.M. Spotted knapweed spread and plant community changes in a lacustrine dune system. Nat. Areas J. 2021, 41, 11–17. [Google Scholar] [CrossRef]
  10. Lichter, J. Colonization constraints during primary succession on coastal Lake Michigan sand dunes. J. Ecol. 2000, 88, 825–839. [Google Scholar] [CrossRef]
  11. Balke, T.; Herman, P.M.J.; Bouma, T.J. Critical transition in disturbance-driven ecosystems: Identifying windows of opportunity for recover. J. Ecol. 2014, 102, 700–708. [Google Scholar] [CrossRef]
  12. Zobel, M.; Öpik, M. Plant and arbuscular mycorrhizal fungal (AMF) communities—Which drives which? J. Veg. Sci. 2014, 25, 1133–1140. [Google Scholar] [CrossRef]
  13. Dech, J.P.; Maun, M.A. Zonation of vegetation along a burial gradient on the leeward slopes of Lake Huron sand dunes. Can. J. Bot. 2005, 83, 227–236. [Google Scholar] [CrossRef]
  14. Rudolph, T.D.; Laidly, P.R. Pinus banksiana Lamb. Jack Pine. In Silvics of North America. Volume 1: Conifers; Burns, R.M., Honkala, B.H., Eds.; Tech. Coords; Agricultural Handbook 654; Forest Service, United States Department of Agriculture: Washington, DC, USA, 1990. Available online: https://www.srs.fs.usda.gov/pubs/misc/ag_654/volume_1/pinus/banksiana.htm (accessed on 30 May 2024).
  15. Chun, Y.M.; Choi, Y.D. Structure, species composition, and soil characteristics in a chronosequence of jack pine (Pinus banksiana Lamb.) stands on the southern shore of Lake Michigan. Am. Midl. Nat. 2012, 168, 408–426. [Google Scholar] [CrossRef]
  16. Abrams, M.D.; Sprugel, D.G.; Dickmann, D.I. Multiple successional pathways on recently disturbed jack pine sites in Michigan. For. Ecol. Manag. 1985, 10, 31–48. [Google Scholar] [CrossRef]
  17. Venier, L.A.; Pearce, J.L. Boreal bird community response to jack pine forest succession. For. Ecol. Manag. 2005, 217, 19–36. [Google Scholar] [CrossRef]
  18. Cayford, J.H.; McRae, D.J. The ecological role of fire in jack pine forests. In Role of Fire in Northern Circumpolar Ecosystems; Wein, R.W., MacLean, D.A., Eds.; John Wiley and Sons: New York, NY, USA, 1983; pp. 183–199. [Google Scholar]
  19. Greene, D.F.; Splawinski, T.B.; Gauthier, S.; Bergeron, Y. Seed abscission schedules and the timing of post-fire salvage of Picea mariana and Pinus banksiana. For. Ecol. Manag. 2013, 303, 20–24. [Google Scholar] [CrossRef]
  20. Pinno, B.D.; Errington, R.C.; Thompson, D.K. Young jack pine and high severity fire combine to create potentially expansive areas of understocked forest. For. Ecol. Manag. 2013, 310, 517–522. [Google Scholar] [CrossRef]
  21. Gauthier, S.; Bergeron, Y.; Simon, J.-P. Effects of fire regime on the serotiny level of jack pine. J. Ecol. 1996, 84, 539–548. [Google Scholar] [CrossRef]
  22. Hernández-Serrano, A.; Verdu, M.; Santos-del-Blanco, L.; Climent, J.; González-Martínez, S.C.; Pausas, J.G. Heritability and quantitative genetic divergence of serotiny, a fire persistence plant trait. Ann. Bot. 2014, 114, 571–577. [Google Scholar] [CrossRef]
  23. Braind, C.H.; Schwilk, D.W.; Gauthier, S.; Bergeron, Y. Does fire regime influence life history traits of jack pine in the southern boreal forest of Québec, Canada? Plant Ecol. 2015, 216, 157–164. [Google Scholar] [CrossRef]
  24. Radeloff, V.C.; Mladenoff, D.J.; Guries, R.P.; Boyce, M.S. Spatial patterns of cone serotiny in Pinus banksiana in relation to fire disturbance. For. Ecol. Manag. 2004, 189, 133–141. [Google Scholar] [CrossRef]
  25. Gauthier, S.; Bergeron, Y.; Simon, J.-P. Cone serotiny in jack pine: Ontogenetic, positional, and environmental effects. Can. J. For. Res. 1993, 23, 394–401. [Google Scholar] [CrossRef]
  26. Luu, H.; Ris Lambers, J.H.; Lutz, J.A.; Metz, M.; Snell, R.S. The importance of regeneration processes on forest diversity in old-growth forests in the Pacific Northwest. Phil. Trans. R. Soc. B 2024, 379, 20230016. [Google Scholar] [CrossRef] [PubMed]
  27. Seidl, R.; Turner, M.G. Post-disturbance reorganization of forest ecosystems in a changing world. Proc. Nat. Acad. Sci. USA 2022, 119, e2202190119. [Google Scholar] [CrossRef] [PubMed]
  28. Sandoval-Martínez, J.; Badano, E.I.; Guerra-Coss, F.A.; Flores Cano, J.A.; Flores, J.; Gelviz-Gelvez, S.M.; Barragán-Torres, F. Selecting tree species to restore forest under climate change conditions: Complementing species distribution models with field experimentation. J. Environ. Manag. 2023, 329, 117038. [Google Scholar] [CrossRef] [PubMed]
  29. Guisan, A.; Weiss, S.B.; Weiss, A.D. GLM versus CCA spatial modeling of plant species distribution. Plant Ecol. 1999, 143, 107–122. [Google Scholar] [CrossRef]
  30. Lissovsky, A.A.; Dudov, S.V. Species-distribution modeling: Advantages and limitations of its application. 2. MaxEnt. Biol. Bull. Rev. 2021, 11, 265–275. [Google Scholar] [CrossRef]
  31. Baldwin, R.A. Use of maximum entropy modeling in wildlife research. Entropy 2009, 11, 854–866. [Google Scholar] [CrossRef]
  32. Phillips, S.J.; Anderson, R.P.; Schapire, R.E. Maximum entropy modeling of species geographic distributions. Ecol. Model. 2006, 190, 231–259. [Google Scholar] [CrossRef]
  33. Phillips, S.J.; Anderson, R.P.; Dudík, M.; Schapire, R.E.; Blair, M.E. Opening the black box: An open-source release of Maxent. Ecography 2017, 40, 887–893. [Google Scholar] [CrossRef]
  34. Harremoës, P.; Topsøe, F. Maximum entropy fundamentals. Entropy 2001, 3, 191–226. [Google Scholar] [CrossRef]
  35. Guiasu, S.; Shenitzer, A. The principle of maximum entropy. Math. Intell. 1985, 7, 42–48. [Google Scholar] [CrossRef]
  36. Bach, D.P. Plant Communities, Habitats, and Soil Conditions of Grand Sable Dunes, Pictured Rocks National Lakeshore, Michigan. Master’s Thesis, Michigan Technological University, Houghton, MI, USA, 1978. [Google Scholar]
  37. Marshall, J.M. Influence of topography, bare sand, and soil pH on the occurrence and distribution of plant species in a lacustrine dune ecosystem. J. Torrey Bot. Soc. 2014, 141, 29–38. [Google Scholar] [CrossRef]
  38. Anderton, J.B.; Loope, W.L. Buried soils in a perched dunefield as indicators of late Holocene lake-level change in the Lake Superior basin. Quat. Res. 1995, 44, 190–199. [Google Scholar] [CrossRef]
  39. Millers, I.; Lachance, D.; Burkman, W.G.; Allen, D.C. North American Sugar Maple Decline Project: Organization and Field Methods; USDA Forest Service General Technical Report NE-154; USDA Forest Service: Radnor, PA, USA, 1991.
  40. US Geological Survey. One Meter x56y517 and x57y517 MI AlgerCo 2015 DEM. Available online: https://www.sciencebase.gov/catalog/item/543e6b86e4b0fd76af69cf4c (accessed on 30 May 2024).
  41. Hop, K.; Lubinski, S.; Dieck, J.; Menard, S.; Drake, J. National Park Service Vegetation Inventory Program: Pictured Rocks National Lakeshore; Natural Resource Report NPS/GLKN/NRR—2010/201; Natural Resource Program Center: Fort Collins, CO, USA, 2010.
  42. US Geological Survey. Lidar Point Cloud—USGS National Map 3DEP Downloadable Data Collection. Available online: https://www.sciencebase.gov/catalog/item/4f70ab64e4b058caae3f8def (accessed on 30 May 2024).
  43. Khosravipour, A.; Skidmore, A.K.; Isenburg, M.; Wang, T.; Hussin, Y.A. Generating pit-free canopy height models from airborne lidar. Photogram. Eng. Remote Sens. 2014, 80, 863–872. [Google Scholar] [CrossRef]
  44. R Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2023; Available online: https://www.R-project.org/ (accessed on 30 May 2024).
  45. Roussel, J.R.; Auty, D.; Coops, N.C.; Tompalski, P.; Goodbody, T.R.H.; Sánchez Meador, A.; Bourdon, J.F.; de Boissieu, F.; Achim, A. lidR: An R package for analysis of Airborne Laser Scanning (ALS) data. Remote Sens. Environ. 2020, 251, 112061. [Google Scholar] [CrossRef]
  46. Roussel, J.R.; Auty, D. Airbore LIDAR Data Manipulation and Visualization for Forestry Applications. R Package Version 4.1.1. Available online: https://cran.r-project.org/package=lidR (accessed on 30 May 2024).
  47. Pebesma, E.J. Multivariable geostatistics in S: The gstat package. Comp. Geosci. 2004, 30, 683–691. [Google Scholar] [CrossRef]
  48. Gräler, B.; Pebesma, E.; Heuvelink, G. Spatio-temporal interpolation using gstat. R J. 2016, 8, 204–218. [Google Scholar] [CrossRef]
  49. Hijmans, R. Terra: Spatial Data Analysis. R Package Version 1.7-39. Available online: https://CRAN.R-project.org/package=terra (accessed on 30 May 2024).
  50. Hijmans, R. Predicts: Spatial Prediction Tools. R Package Version 0.1-8. Available online: https://CRAN.R-project.org/package=predicts (accessed on 30 May 2024).
  51. Freeman, E.A.; Moisen, G.G. A comparison of the performance of threshold criteria for binary classification in terms of predicted prevalence and kappa. Ecol. Model. 2008, 217, 48–58. [Google Scholar] [CrossRef]
  52. Maun, M.A. Colonizing ability of Ammophila breviligulata through vegetative regeneration. J. Ecol. 1984, 72, 565–574. [Google Scholar] [CrossRef]
  53. Voss, E.G. Michigan Flora. Part I. Gymnosperms and Monocots; Bulletin 55; Cranbrook Institute of Science: Bloomfield Hills, MI, USA, 2001. [Google Scholar]
  54. Kashian, D.M.; Sosin, J.R.; Huber, P.W.; Tucker, M.M.; Dombrowski, J. A neutral modeling approach for designing spatially heterogeneous jack pine plantations in northern Lower Michigan, USA. Land. Ecol. 2017, 32, 1117–1131. [Google Scholar] [CrossRef]
  55. Filicetti, A.T.; LaPointe, R.A.; Nielson, S.E. Effects of fire severity and woody debris on tree regeneration for exploratory well pads in jack pine (Pinus banksiana) forests. Forests 2021, 12, 1330. [Google Scholar] [CrossRef]
  56. Whittle, C.A.; Duchesne, L.C.; Needham, T. The impact of broadcast burning and fire severity on species composition and abundance of surface vegetation in a jack pine (Pinus banksiana) clear-cut. For. Ecol. Manag. 1997, 94, 141–148. [Google Scholar] [CrossRef]
  57. Poulson, T.L. Mechanisms of Plant Succession on the Sand Dunes along the Southern Lake Michigan; A research report submitted to the Science Division; National Park Service—Indiana Dunes National Lakeshore: Porter, IN, USA, 1993. [Google Scholar]
  58. Day, M.E.; Schedlbauer, J.L.; Livingston, W.H.; Greenwood, M.S.; White, A.S.; Brissette, J.C. Influence of seedbed, light environment, and elevated night temperature on growth and carbon allocation in pitch pine (Pinus rigida) and jack pine (Pinus banksiana) seedlings. For. Ecol. Manag. 2005, 205, 59–71. [Google Scholar] [CrossRef]
  59. Powers, M.D.; Pregitzer, K.S.; Palik, B.J. Physiological performance of three pine species provides evidence for gap partitioning. For. Ecol. Manag. 2008, 256, 2127–2135. [Google Scholar] [CrossRef]
  60. Deighton, H.D.; Bell, F.W.; Thiffault, N.; Searle, E.B.; Leitch, M.; Sharma, M.; Dacosta, J. Trade-offs among release treatments in jack pine plantations: Twenty-five year responses. Forests 2021, 12, 370. [Google Scholar] [CrossRef]
  61. Montgomery, R.A.; Palik, B.J.; Boyden, S.B.; Reich, P.B. New cohort growth and survival in variable retention harvests of a pine ecosystem in Minnesota, USA. For. Ecol. Manag. 2013, 310, 327–335. [Google Scholar] [CrossRef]
  62. Wilson, L.F. A Guide to Insect Injury of Conifers in the Lake States; Agriculture Handbook 501; Forest Service, United States Department of Agriculture: Washington, DC, USA, 1977.
  63. McCullough, D.G. A review of factors affecting the population dynamics of jack pine budworm (Choristoneura pinus pinus Freeman). Pop. Ecol. 2000, 42, 243–256. [Google Scholar] [CrossRef]
  64. McCullough, D.G.; Kulman, H.M. Differences in foliage quality of young jack pine (Pinus banksiana Lamb.) on burned and clearcut sites: Effects on jack pine budworm (Choristoneura pinus pinus Freeman). Oecologia 1991, 87, 135–145. [Google Scholar] [CrossRef]
  65. Nealis, V.G.; Lomic, P.V. Host-plant influence on the population ecology of the jack pine budworm, Choristoneura pinus (Lepidoptera: Tortricidae). Ecol. Entomol. 1994, 19, 367–373. [Google Scholar] [CrossRef]
  66. Batzer, H.O.; Jennings, D.T. Numerical analysis of a jack pine budworm outbreak in various densities of jack pine. Environ. Entomol. 1980, 9, 514–524. [Google Scholar] [CrossRef]
  67. Latsch, M.E. Effects of Management on Native and Exotic Plant Communities in Pictured Rocks National Lakeshore in the Upper Peninsula of Michigan. Ph.D. Thesis, Michigan Technological University, Houghton, MI, USA, 2011. [Google Scholar]
  68. Grzesiak, K.E. Long-Term Effects of Herbicide Treatments on Spotted Knapweed and Non-Target Plants in the Grand Sable Dunes. Master’s Thesis, University of Michigan, Ann Arbor, MI, USA, 2013. [Google Scholar]
Forests 15 01138 g001
Figure 1. Grand Sable Dunes survey area aerial image with dune boundary (white line) with inset of site location in the United States (gray star) (A), 2018 transect locations (B), 2020 plot locations (C), 2021 plot locations (D), and 2022 plot locations (E).
Figure 1. Grand Sable Dunes survey area aerial image with dune boundary (white line) with inset of site location in the United States (gray star) (A), 2018 transect locations (B), 2020 plot locations (C), 2021 plot locations (D), and 2022 plot locations (E).
Forests 15 01138 g001
Forests 15 01138 g002
Figure 2. Frequency of Pinus banksiana Lamb. seedlings (≤5 mm root collar diameter) and saplings (>5 mm, ≤15 mm root collar diameter) between years in 1 mm size classes.
Figure 2. Frequency of Pinus banksiana Lamb. seedlings (≤5 mm root collar diameter) and saplings (>5 mm, ≤15 mm root collar diameter) between years in 1 mm size classes.
Forests 15 01138 g002
Forests 15 01138 g003
Figure 3. Relative abundance for the top five most abundant seedlings (A) and sapling (B) species pooled across years (2020–2022). ‘Other’ includes 22 additional species for seedlings and 21 for saplings.
Figure 3. Relative abundance for the top five most abundant seedlings (A) and sapling (B) species pooled across years (2020–2022). ‘Other’ includes 22 additional species for seedlings and 21 for saplings.
Forests 15 01138 g003
Forests 15 01138 g004
Figure 4. Partial response probabilities for Pinus banksiana Lamb. seedling (A–D) and sapling (E–H) models for each independent variable. For basal area (B,F), solid lines represent P. banksiana basal area and dashed lines represent basal area pooled for all other species. Vegetation types are categorical numerals from the original data.
Figure 4. Partial response probabilities for Pinus banksiana Lamb. seedling (A–D) and sapling (E–H) models for each independent variable. For basal area (B,F), solid lines represent P. banksiana basal area and dashed lines represent basal area pooled for all other species. Vegetation types are categorical numerals from the original data.
Forests 15 01138 g004
Forests 15 01138 g005
Figure 5. Probability map for Pinus banksiana Lamb. seedlings (A) and saplings (B) as predicted maximum entropy models based on aspect, other species basal area, P. banksiana basal area, slope, and vegetation types. Minimum of z-range set to maximum kappa from evaluation with model testing data.
Figure 5. Probability map for Pinus banksiana Lamb. seedlings (A) and saplings (B) as predicted maximum entropy models based on aspect, other species basal area, P. banksiana basal area, slope, and vegetation types. Minimum of z-range set to maximum kappa from evaluation with model testing data.
Forests 15 01138 g005
Forests 15 01138 g006
Figure 6. Probability of Pinus banksiana seedling occurrence at true positive locations within vegetation cover types.
Figure 6. Probability of Pinus banksiana seedling occurrence at true positive locations within vegetation cover types.
Forests 15 01138 g006
Forests 15 01138 g007
Figure 7. Probability of Pinus banksiana sapling occurrence at true positive locations within vegetation cover types.
Figure 7. Probability of Pinus banksiana sapling occurrence at true positive locations within vegetation cover types.
Forests 15 01138 g007
Table 1. Number of plots surveyed, total seedling and sapling abundance, and Pinus banksiana Lamb. seedling and sapling abundance by year.
Table 1. Number of plots surveyed, total seedling and sapling abundance, and Pinus banksiana Lamb. seedling and sapling abundance by year.
TotalPinus banksiana
YearPlotsSeedlingsSaplingsSeedlingsSaplings
20205088325824
20211742403958235187
202211796410009895
Table 2. Seedling and sapling maximum entropy model variable contribution and importance.
Table 2. Seedling and sapling maximum entropy model variable contribution and importance.
Seedling ModelSapling Model
VariablePercent
Contribution		Permutation
Importance		Percent
Contribution		Permutation
Importance
Aspect7.25.315.515.9
Other Species Basal Area28.237.99.114.3
Pinus banksiana Basal Area14.319.042.546.9
Slope24.918.56.37.7
Vegetation Type25.419.326.615.2
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Danielson, J.C.; Warrix, A.R.; Lehman, M.E.; Lehman, A.C.; Marshall, J.M. Pinus banksiana Lamb. Regeneration Patterns in a Lacustrine Dune System. Forests 2024, 15, 1138. https://doi.org/10.3390/f15071138

AMA Style

Danielson JC, Warrix AR, Lehman ME, Lehman AC, Marshall JM. Pinus banksiana Lamb. Regeneration Patterns in a Lacustrine Dune System. Forests. 2024; 15(7):1138. https://doi.org/10.3390/f15071138

Chicago/Turabian Style

Danielson, Jonathan C., Adam R. Warrix, Madison E. Lehman, Andrew C. Lehman, and Jordan M. Marshall. 2024. "Pinus banksiana Lamb. Regeneration Patterns in a Lacustrine Dune System" Forests 15, no. 7: 1138. https://doi.org/10.3390/f15071138

APA Style

Danielson, J. C., Warrix, A. R., Lehman, M. E., Lehman, A. C., & Marshall, J. M. (2024). Pinus banksiana Lamb. Regeneration Patterns in a Lacustrine Dune System. Forests, 15(7), 1138. https://doi.org/10.3390/f15071138

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop