Home

Background:

Biodiversity underpins ecosystem processes and services, described in policy terms as nature's contributions to people (NCP). Agricultural land conversion is a leading driver of biodiversity loss, yet how it affects multiple NCPs via direct and cascading indirect effects remains poorly understood. Here, we develop a network-based approach to quantify direct and indirect effects of land conversion on multiple NCPs, using an empirical multi-interaction network from an organic farm in Southwest England. Simulating a sequential transition from extensive organic to intensive non-organic management, we assess how species and interaction loss shape seven NCPs. Cascading indirect effects were the primary drivers of inferred NCP change, with indirect losses (97%) three-fold greater than direct losses (31%) across the gradient. Effects varied among NCPs: pollination declined by nearly 95%, while total crop biomass rose by up to 191%, revealing a trade-off between production and other benefits such as pollination and cultural services. Null model analyses show that NCP losses are driven disproportionately by species identity rather than richness alone. These findings demonstrate that indirect species interactions are central to sustaining multifunctional agroecosystems, and that their erosion under land conversion amplifies trade-offs among NCPs, underscoring the need to incorporate community structure into land management strategies.

---

📁 Repository structure

├── 🧑‍💻 R/          All code used in the project
├── 📊 Data/       Processed outputs, raw data (Raw_data/), and external repository data (Repositories/)
└── 🖼️ Graphs/     Every figure produced by the project

---

🧰 Utility

Script	What it does
⚙️ functions.R	Shared functions used across more than one analysis script. Functions are grouped by purpose, each with a short explanation above it.

---

🔄 Data processing

Run in order — each step feeds into the next.

Script	What it does	📥 Inputs	📤 Outputs
1_Abundances.R	Estimates species abundances and builds the species and interaction lists from raw field data.	Raw_data/PLANT_HABITAT3.csv Raw_data/nore2.csv Repositories/Orford_et_al_2015.csv Repositories/DoPI.csv	species_abundances.csv elist_nore.csv nodes.csv pollinators_sp.csv
2_Multi_object.R	Organizes everything into a multilayer network object: nodes, layers (habitats), interactions, and species-per-habitat.	elist_nore.csv nodes.csv pollinators_sp.csv Raw_data/service_edgelist.csv	Norwood_farm.RData
3_Biomass.R	Assigns biomass to species using databases and literature (see Appendix S1 for sources).	Norwood_farm.RData Raw_data/nore2.csv Repositories/BirdFuncDat.txt Repositories/MamFuncDat.txt Repositories/bodysizes_2008.txt Repositories/coleoptera.xlsx	biomass.csv

---

📈 Main analysis

Outputs listed below are the main files only — see the Data for the complete list, including intermediate files.

Script	What it does	📥 Inputs	📤 Outputs
4_Land_conversion_simulation.R	Simulates land conversion from extensive organic to intensive non-organic management, and estimates NCP provision + proportion retained.	species_abundances.csv Norwood_farm.RData biomass.csv Raw_data/habitatarea.csv	Land_use_state_nodes_M1_M2.csv Land_use_edgelist_M1_M2.csv Land_use_dir_ES_M1_M2.csv Land_use_ind_ES_M1_M2.csv
5_Null_model.R	Builds a null model that removes the same number of species as the real simulation, but at random, and calculates Z-scores against it.	Norwood_farm.RData species_abundances.csv biomass.csv Raw_data/habitatarea.csv Land_use_state_nodes_M1_M2.csv Land_use_edgelist_M1_M2.csv Land_use_dir_ES_M1_M2.csv Land_use_ind_ES_M1_M2.csv	state_node_sim_CP_M1_M2_A2.csv edge_list_sim_CP_M1_M2_A2.csv z_score_dir_ES_CP_M1_M2_A2.csv z_score_amount_ES_CP_M1_M2_A2.csv z_score_ind_ES_CP_M1_M2_A2.csv
6_Sps_indirect_contribution.R	Estimates each species' indirect contribution to NCP provision per scenario, and tests how land conversion affects it.	Norwood_farm.RData Land_use_edgelist_M1_M2.csv	Land_use_shortpath_M1_M2.csv
7_Graph_networks.R	Produces Figure 2: network plots for the E, M, and IN management scenarios.	Norwood_farm.RData Land_use_edgelist_M1_M2.csv	Network plot PDFs in Graphs/
8_Graphs.R	Produces Figures 3–5 and the supplementary figures.	Land_use_dir_ES_M1_M2.csv Land_use_ind_ES_M1_M2.csv Land_use_state_nodes_M1_M2.csv Land_use_shortpath_M1_M2.csv z_score_dir_ES_CP_M1_M2_A2.csv z_score_amount_ES_CP_M1_M2_A2.csv z_score_ind_ES_CP_M1_M2_A2.csv Norwood_farm.RData	Figure PDFs in Graphs/