Skip to main content
Dryad logo

Optimising nature conservation outcomes for a given region-wide level of food production

Citation

Finch, Tom et al. (2020), Optimising nature conservation outcomes for a given region-wide level of food production, Dryad, Dataset, https://doi.org/10.5061/dryad.69p8cz8xs

Abstract

  1. The land sharing-sparing framework aims to quantify the trade-off between food production and biodiversity conservation, but it has been criticised for offering, for reasons of simplicity, an unrealistically limited set of different land uses.
  2. Here, we develop the framework to evaluate a much larger suite of land-use strategies in which the areas and yields of three land-use compartments, natural habitat, high-yield farmland, and lower-yield farmland, are varied simultaneously. For two regions of England, we use functions that relate the local population density of breeding bird species to farm yield to simulate species-specific region-wide population sizes under each strategy.
  3. We find that geometric mean population size, averaged across all species, is maximised when farmland yields are higher than the current average yield, and when spared land combines both natural habitat and low-yield farmland. This conclusion is relatively insensitive to the maximum yield considered feasible under extreme land sparing, and holds across a range of region-wide food production targets. 
  4. To some extent, our conclusion depends on which species groups were included in the assessment. Considered alone, farmland birds preferred a strategy with little or no natural habitat. Nonetheless, the optimal strategy was broadly consistent across all widely-recognised listings of species of conservation concern.
  5. Synthesis and applications. Our study looks at landscapes with a long history of agricultural use, and which have little remaining natural habitat, such as those found in lowland England. In these landscapes, conservation outcomes are likely to be maximised by mixed strategies in which high-yield production enables an increase in area of both natural habitat and low-yield farmland elsewhere in the region. 

Usage Notes

We have provided annotated R script (optimal_strategy_script.R) along with 7 input files.

This code creates scenarios and calculates geometric mean population size, allowing all figures (1-5) to be reproduced.

 

Input file 1: current.csv

Current (2015) properties of each 1-km square in each region

gridref Unique identifier for each 1-km square

nca National Character Area (either The Fens or Salisbury Plain)

habitat Either farmed, or wet grassland or fen in The Fens or chalk grassland or woodland in Salisbury Plain

unbuilt Proportion of square which is not urban or suburban

currency Yield currency (energy)

yield Yield of unbuilt land (GJ per ha)

production Production from all land (yield * unbuilt)

 

Input file 2: spared.csv

Average properties of spared land

habitat wet grassland or fen in The Fens or chalk grassland or woodland in Salisbury Plain

currency Yield currency (energy)

unbuilt Total area which is not urban or suburban

yield Average yield (GJ per ha)

production Average production (yield)

nca National Character Area (either The Fens or Salisbury Plain)

 

Input file 3: dy_pars.csv

Density-yield parameters

currency Yield currency (energy)

spp 2-letter species code

english Species name

max.density Maximum observed density

mean.density Mean observed density

max.yield Maximum yield across survey sites

upwards TRUE if density-yield curve trends upwards at max.yield, otherwise FALSE

mod A if 3-parameter model, B if 4-parameter model

b0 Density-yield parameter (see equations 1 and 2 in Finch et al. (2019) DOI: 10.1111/cobi.13316)

b1 Density-yield parameter (see equations 1 and 2 in Finch et al. (2019) DOI: 10.1111/cobi.13316)

b2 Density-yield parameter (see equations 1 and 2 in Finch et al. (2019) DOI: 10.1111/cobi.13316)

alpha Density-yield parameter (see equations 1 and 2 in Finch et al. (2019) DOI: 10.1111/cobi.13316)

top Top model (A or B)

nca National Character Area (either The Fens or Salisbury Plain)

 

Input file 4: traits.csv

Species traits (each species appears in multiple rows, according to group identity)

spp 2-letter species code

english Species name

nca National Character Area (either The Fens or Salisbury Plain)

group Group - see main text. 

 

Input file 5: wg_fens.csv

Density in wet grassland (for The Fens only)

spp 2-letter species code

english Species name

habitat wet grassland

density Average density

 

Input file 6: wd_sali.csv

Density in woodland (for Salisbury Plain only)

spp 2-letter species code

english Species name

habitat woodland

density Average density

 

Input file 7: fenwg_fens.csv

Additional natural habitat specialists (for The Fens only)

spp 2-letter species code

english Species name

habitat fen or wet grassland

density Relative density 1 (present) or 0 (absent)