Insects and non-woody plants slow down tropical forest succession: a community-wide experiment in Papua New Guinea
Data files
Apr 15, 2026 version files 309.73 KB
-
alien_native_plants_WP_rev2.R
14.21 KB
-
Biomass_NWP_WP_combine_rev2.R
17.32 KB
-
Combine_Sites_Biomass_2023.csv
181.98 KB
-
CombineSite_Insects_Orders_data.csv
24.08 KB
-
Diversity_NWP_WP_combine_rev2.R
19.76 KB
-
Insect_abundance_in_plots_rev2.R
13.16 KB
-
NWP_WP_RDA.R
32.74 KB
-
README.md
6.48 KB
Abstract
Ecological succession is a complex community-wide process whose theoretical principles require further development. Here, we test the hypothesis that the succession of woody plants in tropical rainforests is determined by bottom-up plant competition, rendering top-down control by insect herbivores insignificant.
Over 18 months of rainforest secondary succession, we removed insects and non-woody plants from replicate 5×5 m plots in a factorial experiment at 700 and 1700 m elevation in New Guinea.
At 700 m elevation, insect removal increased biomass, reduced diversity, and altered the species composition of woody plants, while removal of non-woody plants increased both biomass and diversity of woody plants and altered their species composition. At 1700 m, the effect of insects on woody plant biomass and species composition disappeared while the effect of non-woody plants on woody plant biomass, diversity and species composition became stronger.
The removal of insects did not increase the proportion of alien species in the woody vegetation as predicted by the enemy-release hypothesis. The increased disturbance caused by removing non-woody plants also did not promote alien plants.
Synthesis. The importance of top-down insect herbivory decreased, while the importance of bottom-up plant competition with non-woody plant species increased with elevation, representing a gradient of increasing environmental stress, in the succession of woody plants.
Metadata for the MS Analyses
This is the dataset and R codes for the MS titled: "Insects and non-woody plants slow down tropical forest succession: a community-wide experiment in Papua New Guinea".
Kari Sogera Iamba, Piotr Szefer, Kenneth Molem, Austin Sau, Gibson Maiah, Vojtech Novotny
(A) Datasets
Two datasets are presented in two Excel files, used for all the analyses:
[1] Plant data from the experimental plots in Combine_Sites_Biomass_2023.xlsx used for all analyses in the main MS and also for the Fig. S1 in the Supporting Information. Each row in the table represents a record of a particular plant species in a particular experimental plot, with information on growth form and the amount of plant biomass. Definition of columns:
(A) Elevation: elevation of the study site, 700 or 1700 m asl.
(B) Blocks: nine blocks of four experimental plots per elevation, B1 – B9 for 700 m asl, B10 - B18 for 1700 m asl, each of them in a 0.2 ha forest clearing. These unique blocks allows for direct use of Blocks as random factor in the linear mixed effects models (LMM) without any need for nesting.
(C) RDA_Blocks: nine blocks of four experimental plots, B1 – B9, used only for RDA ordination since separate analyses was done for non-woody and woody plants for 700 m and 1700 m asl.
(D) Treatments: experimental treatment (C: control, I: insecticide, W: weeding, WI: weeding + insecticide). There was one plot per block for each treatment, so that the combination of columns A, B, and C defines each of the 72 plots.
(E) Plant_sp: the name of the species or, for unidentified species, a species code
(F) Family: plant family
(G) Plants: classification of each species as woody or non-woody
(H) Growth_Form: classification of each species into fern, herb, grass, sedge, shrub, liana , palm or tree growth form
(I) Status: classification of each species as native to New Guinea or alien
(J) Biomass_kg: fresh plant biomass at the end of the experiment for each species
[2] Arthropod data from the experimental plots in CombineSite_Insects_Orders_data.xlsx used for analyses in the Supporting Information. Each row in the table represents a record of the number of individuals from a particular arthropod order found at the end of the experiment in each plot. Definition of columns:
(A) Elevation: elevation of the study site, 700 or 1700 m asl.
(B) Blocks: nine blocks of four experimental plots per elevation, B1 – B9 for 700 m asl, B10 - B18 for 1700 m asl, each of them in a 0.2 ha forest clearing. These unique blocks allows for direct use of Blocks as random factor in the linear mixed effects models (LMM) without any need for nesting.
(C) Treatments: experimental treatment (C: control, I: insecticide, W: weeding, WI: weeding + insecticide). There was one plot per block for each treatment, so that the combination of columns A, B, and C defines each of the 72 plots.
(D) Order: defines the arthropod order
(E) Guilds: classified arthropods into herbivores and other guilds
(F) Abundance: the number of individuals found in each plot at the end of the experiment
(B) R code files
Use the read_excel function from library(readxl) to load the data. The data are located in the folder called "datasets". For example: read_excel("datasets/Combine_Sites_Biomass_2023.xlsx", sheet="combine.site.biomass") [Combine_Sites_Biomass_2023 is a excel file located in datasets folder with sheet name "combine.site.biomass"]
The repository consists of seven script files used for the analyses. Five scripts (#1-5) were used for the main manuscript and supporting information:
(1) Insect_abundance_in_plots_rev2.R: scripts for all insect abundance and herbivore abundance corresponding to each treatment [C, I, W, WI].
(2) Biomass_NWP_WP_combine_rev2.R: scripts for analyzing biomass of non-woody plants (NWP), woody plants (WP) and their combined biomasses corresponding to each treatment [C, I, W, WI].
(3) Diversity_NWP_WP_combine_rev2.R: scripts for analyzing divesity of non-woody plants (NWP), woody plants (WP) and their combined biomasses corresponding to each treatment [C, I, W, WI].
(4) NWP_WP_RDA.R: scripts for analyzing diversity of non-woody plants (NWP) and woody plants (WP) species composition based on biomasses corresponding to each treatment [C, I, W, WI].
(5) alien_native_plants_WP_rev2.R: scripts for analyzing biomass of alien and native plants based on woody plant biomass corresponding to each treatment [C, I, W, WI].
(C) Information on statistical tables (in Supporting Information section)
(1) Statistical table: TABLE S1 shows the ANOVA output of LMM models based on insect herbivore abundance and others (non herbivorous insects and spiders).
(2) Statistical table: TABLE S2 shows the summary output for the main and interaction effect of insecticide, weeding and elevation on insect herbivore and others (non herbivorous insects and spiders).
(3) Statistical tables:TABLE S3 shows the ANOVA output of LMM models based on non-woody, woody and total plant biomass.
(4) Statistical table: TABLE S4 shows the summary output for the main and interaction effect of insecticide, weeding and elevation on non-woody, woody and total plant biomass.
(5) Statistical tables:TABLE S5 shows the ANOVA output of LMM models based on non-woody, woody and total plant diversity.
(6) Statistical table: TABLE S6 shows the summary output for the main and interaction effect of insecticide, weeding and elevation on non-woody, woody and total plant diversity.
(7) Statistical table: TABLE S7 (RDA ordination) is extracted by the multiconstrained function in BiodiversityR package: multiconstrained(method = "rda", formula = species.matrix ~ Treatments + Condition(Blocks), data = data, scaling=2, distance = "bray").
(8) Statistical table: TABLE S8 (species scores and associated squared correlations) is extracted from RDA models using the function envfit(plot, env=species.matrix).
(9) Statistical table: TABLE 9 shows the LMM model ANOVA output of alien WP proportion based on biomass and richness. Proportion is calculated using the logit proportion using this formula: logit(p) = log(p / 1 - p), where p = alien proportion and 1 − p = native proportion.
(10) Statistical table: TABLE 10 shows the LMM model summary output of alien WP proportion based on biomass and richness. Proportion is calculated using the logit proportion using this formula: logit(p) = log(p / 1 - p), where p = alien proportion and 1 − p = native proportion.