Active restoration increases tree species richness and recruitment of large-seeded taxa after 16-18 years
Data files
Sep 25, 2024 version files 10.15 MB
-
ECOLAPPS_final_Schubert.et.al.2024.csv
10.14 MB
-
PlotQuadrantes.csv
348 B
-
README.md
10.03 KB
Abstract
Tropical forest restoration presents a potential lifeline to mitigate climate change and biodiversity crises in the Anthropocene. Yet, the extent to which human interventions, such as tree planting, accelerate the recovery of mature functioning ecosystems or redirect successional trajectories towards novel states remains uncertain due to a lack of long-term experiments. In 2004-2006, we established three 0.25-ha plots at ten sites in southern Costa Rica to test three forest restoration approaches: natural regeneration (no planting), applied nucleation (planting in patches), and plantation (full planting). In a comprehensive survey after 16-18 years of recovery, we censused >80,000 seedlings, saplings, and trees across 26 restoration plots (9 natural regeneration, 9 applied nucleation, 8 plantation) and six adjacent reference forests to evaluate treatment effects on recruitment patterns and community composition. Both applied nucleation and plantation treatments resulted in significantly elevated seedling and sapling establishment and more predictable community composition compared to natural regeneration. Similarity of vegetation composition to reference forest tended to scale positively with treatment planting intensity. Later-successional species with seeds ≥5 mm had significantly greater seedling and sapling abundance in the two planted treatments, and plantation showed similar recruitment densities of large-seeded (≥10 mm) species to reference forest. Plantation tended towards a lower abundance of early-successional recruits than applied nucleation. Trees (≥5 cm DBH) in all restoration treatments continued to be dominated by a few early-successional species and originally transplanted individuals. Seedling recruits of planted taxa were more abundant in applied nucleation than the other treatments though few transitioned into the sapling layer. Overall, our findings show that active tree planting accelerates the establishment of later-successional trees compared to natural regeneration after nearly two decades. While the apparent advantages of higher density tree planting on dispersal and understory establishment of larger-seeded, later-successional species recruitment is notable, more time is needed to assess whether these differences will persist and transition to the more rapid development of a mature later-successional canopy. Results underscore the need for ecological restoration planning and monitoring that targets biodiversity recovery over multiple decades.
README: Active restoration increases tree species richness and recruitment of large-seeded taxa after 16-18 years
https://doi.org/10.5061/dryad.s7h44j1gc
Description of the data and file structure
STUDY AREA AND BACKGROUND ON EXPERIMENT
Data collection occurred Jun-Jul 2022 in Coto Brus county of southern Costa Rica near Las Cruces Biological Station. The data represent full censuses of 26 plots from 10 restoration sites and six plots sampled from mature reference forests in adjacent areas at six of these locations. Restored were designed to have an area of 0.25 ha with either 48x48 m or 42x54 m dimensions with a ≥5-m buffer. Restoration experiments initiated in 2004-2006, when each plot received one of three randomized treatments: natural regeneration n = 9 plots, applied nucleation n = 9, or plantation n = 8. Some of the sites were missing one or two treatments due to their being destroyed or heavily damaged by this point in the study and so were not censused. Plantations were uniformly planted with tree seedlings. Applied nucleation treatment was planted with six tree islands of three sizes: two each of 4×4, 8×8 and 12×12 m. Planted seedling spacing was kept constant (~2.8 m) in plantation and applied nucleation treatments; 313 trees were planted in plantation, 86 in applied nucleation, and none in natural regeneration plots (see Holl et al. 2011 for further details). All plots (including natural regeneration) were cleared to ground level by machete at ~3-mo intervals for the first 2.5 yr to allow planted tree seedlings to grow above existing vegetation. We planted seedlings (20-30 cm tall) of four tree species; these included two native late-successional species, Terminalia amazonia (J.F. Gmel.) Exell (Combretaceae) and Vochysia guatemalensis Donn. Sm. (Vochysiaceae), and two naturalized early-successional species, Erythrina poeppigiana (Walp.) Skeels and Inga edulis Mart. (both Fabaceae) that are used widely in intercropping systems in Central America. By the time of the surveys presented here, a majority of the planted softwoods had died (mean survival and standard deviation: E. poeppigiana 34.5 ± 28.5%; I. edulis 22.9 ± 18.5%). Survival of the other two species remained high (V. guatemalensis 82.8 ± 18.4%; T. amazonia 82.1% ± 17.8%). At six sites, we also sampled adjacent remnant forests to serve as references. These reference forests ranged in size from 2-320 ha and showed no evidence of clearing in the last 75 years since aerial photographs have been available (Zahawi et al. 2015b). All have all been impacted to some extent by human disturbance, as “pristine” forests are not present in our study area (Clement and Horn 2001).
DATA COLLECTION
16-18 years after establishing plots, we censused all naturally recruiting trees ≥20 cm height (hereafter “recruits”) throughout each restoration treatment plot. We mapped individual recruits to a grid of 3×3 m quadrats and counted the number of seedlings >=20 cm but <1 cm diameter at breast height (DBH) of each species within each quadrat. We measured DBH for all recruits >=1 cm and categorized each into sapling (1-<5 cm DBH) or tree (>=5 cm DBH) size classes. If an individual had multiple stems, we based our size classification on the largest stem. Individual restoration plots were mostly 48×48 m (256 3×3 m quadrats), but in some cases plots were 42×54 m (252 quadrats) due to constraints in the landscape at initial plot set up. In a few cases, a smaller plot area was sampled primarily due to extensive anthropogenic damage to a section of the plot (Table S1). The layout of the reference forest plots was slightly different and consisted of four 21×21 m plots (196 quadrats) at five sites and three 21×21 m plots (147 quadrats) at one site. Tree recruits were identified to the lowest operational taxonomic unit following the nomenclature of Tropicos (https://www.tropicos.org).
Citations:
Clement, R. M., and S. P. Horn. 2001. Pre-Columbian land-use history in Costa Rica: a 3000-year record of forest clearance, agriculture and fires from Laguna Zoncho. The Holocene 11:419–426.
Holl, K. D., R. A. Zahawi, R. J. Cole, R. Ostertag, and S. Cordell. 2011. Planting seedlings in tree islands versus plantations as a large-scale tropical forest restoration strategy. Restoration Ecology 19:470–479.
Zahawi, R. A., G. Duran, and U. Kormann. 2015b. Sixty-seven years of land-use change in southern Costa Rica. PloS One 10:e0143554.
Files and variables
File: ECOLAPPS_final_Schubert.et.al.2024.csv
Description: The data presented here have been redacted slightly from the original survey spreadsheets. Namely, we have removed columns for data that were recorded for purposes beyond the scope of the manuscript (e.g., presence/cover of herbaceous plants, plaque numbers of trees part of smaller permanently monitored quadrats, descriptive comment columns). Several calculated columns have been added to classify stem sizes based on maximum diameter. Furthermore, our inclusion of data from planted tree species comes from a separate, distinct survey protocol to document survival of the original 4 planted species. This latter protocol documented survival of all planted stems, but measured diameter from only a subset. For this reason, we have extrapolated size estimates for un-measured stems such that a fixed diameter measurement is assigned (i.e., 0, 1, 5). The assignment was made such to preserve the relative ratio of seedling, sapling, and tree size classes for each plot-species based on the empirical measurements. These extrapolations were made for approximately 35% of the planted stems in plantation plots and less than 1% in applied nucleation plots.
Persons responsible for data entry and proofing
Botanical Identification Person Initials: ARO,FOB,JAR,JF,KDH,RAZ
Data Entry/Proofing Person Initials: BR,CB,FOB,JOP,KDH,MSJ,SCS,THL,VRM
Number of variables: 18
Number of cases/rows: 67046
Variables
- Especie: String; taxonomic codes for each distinguishable taxonomic unit, "ND" denotes "Quadrante" sampling units where no stems were present
- Sitio: String; two-letter site codes corresponding to 10 locations described in Table S1 of Schubert et al. 2024
- Trat: String; two letter codes indicating forest type or treatment, NR - Natural Regeneration, AN - Applied Nucleation, PL - Plantation, RF - Reference Forest
- Parcela: Numeric; numbers from 1-78 indicating the 6x6 m grid cell from each restoration plot. Standard 48x48 m plots have parcela numbers up to 64. Rectangular 42x54 m have numbers up to 78
- Quadrante: String; letter codes for the four subdivisions (each 3x3 m) of a 6x6 m grid cell
- No..indiv.: Numeric; recorded number of individuals of each species each class in each 3x3 "Quadrante" sampling unit
- Altura: Numeric; height in centimeters of a stem reported "seedling" stems that do not reach 1 cm in diameter at breast height, values represent the medians of binned categories from the census protocol where 35 = 20-<50 cm, 75 = 50-<100 cm, 125 = 100-<150 cm, 175 = 150-<200 cm, 201 = >=200 cm, larger size classes are recorded as NA
- DAPmax: Numeric; stem diameter at breast height in centimeters, representing the largest stem in cases where multiple stems corresponded to a single individual
- Origin: String; code indicating the origin of a recorded individual distinguishing between: Recruit - naturally established individual, Planted - an individual that was originally planted in a restoration plot
- Measured: String; code indicating how individual measurements were produced: Empirical - direct measurement of individuals during the census, Extrapolated - applies to a subset of surviving planted stems whose approximate size was estimated based on a subset of individuals measured empirically
- Class: String; description distinguishing the three size classes of recorded individuals: seedling - individuals with stems <1 cm, sapling - individuals with stems 1-<5 cm, tree - individuals with stems >=5 cm
- Family: String; family classifcation following the nomenclature of Tropicos (https://www.tropicos.org)
- Genus: String; genus classifcation following the nomenclature of Tropicos (https://www.tropicos.org), NA is given for taxa identified only to family level
- Specific.Epithet: String; specific epithet classifcation following the nomenclature of Tropicos (https://www.tropicos.org), NA is given for taxa identified to only family or genus level
- Authority: String; naming authority following the nomenclature of Tropicos (https://www.tropicos.org), NA is given for taxa identified to only family or genus level
- disp_mode: String; codes for dispersal mode, ZOO - zoochory, GRAV - gravity, WIND - anemochory, EXPL - ballistic
- succ_stage: String; codes for successional stage at which taxa are most represented, Early - early successional, Mid - mid successional, Late - late successional, UNK - successional stage uncertain
- seed_size: String; numeric codes for seed size based on width, 1 - seeds <5 mm, 2 - 5-<10 mm, 3 - >=10 mm
File: PlotQuadrantes.csv
Description: A supplementary data file used to calculate sampled area in each plot based on the total number of 3x3 m quadrats.
Number of variables: 2
Number of cases/rows: 33
Variables
- Site.Treat: String; codes for all 33 of the sample plots, first two letters are derived from "Sitio" in the main file, second two letters are derived from "Trat" in the main file
- N.Quad: Total number of 3x3 m "Quadrante" (main file) sample units recorded from each plot
File: FinalAnalysis_ECOLAPP_Schubert.et.al.R
Description: R code to reproduce the figures and statistical analyses in the manuscript as well as tables/figures in the supporting information section.