Appendix K: Effects of management on litter distribution (Experiment 1)
The probability that a cell in a given stratum was occupied by litter was modelled at the plot level in a binomial trials GLM. For the analysis, the experimental unit was taken to be all the sample cells in the same stratum in the plot. The response variable was the number of sample cells in that stratum that contained litter and the denominator was the number of cells in that stratum containing any vegetation. Over-dispersion was marked and this was controlled by scaling the model deviance by the dispersion factor (Pearson Chi-squared / residual degrees of freedom). Farm was fitted as a fixed effect.
1. Amount of litter removed by harrowing
Litter measurements were taken before and after harrowing in 2007 and 2008. Using the modelling approach described above, there were no detectable differences between litter densities before and after harrowing (P(before-after) > 0.98 and P(before-after*stratum) > 0.31, for lenient and moderate treatments). Small, non-significant trends were observed: litter densities increased after harrowing in the lower strata (0-5cm and 5-10cm) but decreased in the higher strata. While large clumps of litter were seen to be removed by the harrows, the harrowing appeared to spread the remaining litter throughout the lower sward strata, rather than reducing its distribution.
2. Effect of experimental treatments on long term changes in litter densities
Litter densities in all strata increased over time (P < 0.0001). There was a significant interaction between year and treatment (P = 0.0002). The rates of increase of litter densities did not differ between lenient and moderate plots (P = 0.98) and both were significantly higher than that on control plots (P < 0.0001). The rate of increase on control plots was low and did not differ significantly from zero (P = 0.9). There was no interaction between stratum and year (P = 0.23). By 2007, after one year of the treatments, litter density was already significantly greater on the lenient and moderate plots, compared to the controls (P < 0.0001), but did not differ between the lenient and moderate plots (P = 0.98). Litter densities on the early closure treatments increased substantially between 2007 and 2009 (Figure K2.1).
Figure K2.1: Changes in litter density (probability that a sample cell in each height stratum contained litter) after one year (2007) and four years (2009) of the experimental treatments. Predicted values from the model described in section 2 are plotted for (a0 Control plots and (b) Lenient plots.
3. Effects of harrow type on long-term changes in litter densities
The type of harrow used did not affect the rate of change of litter densities on either the lenient treatment (P(harrow type*year) = 0.95) or the moderate treatment (P(harrow type*year) = 0.72). This model compared the use of spring-tine harrows, chain harrows and two farms that did not harrow, suggesting that harrowing had little effect on the rates of litter accumulation.
10
Appendix L: Summary statistics for bird usage
Table L1: Bird usage on Experiment 1 sites during May-September. Cumulative totals of foraging visits by each species are tabulated for each treatment, closure period (pre- and post-closure) and year. Species are sorted in descending order of overall usage.
grouping / species/guild / Lenient / Moderate / Control / PRE / POST / 2006 / 2007 / 2008 / 2009species / Swallow / 892 / 956 / 1269 / 1039 / 2078 / 596 / 1272 / 700 / 549
House martin / 258 / 380 / 189 / 492 / 335 / 72 / 257 / 304 / 194
Goldfinch / 72 / 203 / 309 / 118 / 466 / 214 / 276 / 55 / 39
Skylark / 243 / 103 / 42 / 233 / 155 / 25 / 114 / 63 / 186
Blackbird / 92 / 104 / 105 / 248 / 53 / 7 / 40 / 132 / 122
Robin / 101 / 119 / 75 / 210 / 85 / 8 / 70 / 149 / 68
Meadow pipit / 110 / 45 / 78 / 7 / 226 / 122 / 101 / 4 / 6
Linnet / 24 / 53 / 49 / 35 / 91 / 2 / 52 / 15 / 57
Magpie / 24 / 57 / 26 / 75 / 32 / 4 / 21 / 52 / 30
Jackdaw / 4 / 88 / 13 / 9 / 96 / 2 / 94 / 4 / 5
Swift / 32 / 34 / 20 / 68 / 18 / 12 / 30 / 44
Rook / 18 / 16 / 49 / 26 / 57 / 15 / 56 / 11 / 1
Wheatear / 14 / 25 / 25 / 64 / 5 / 41 / 13 / 5
Starling / 17 / 12 / 19 / 25 / 23 / 21 / 12 / 15
Carrion crow / 8 / 21 / 11 / 22 / 18 / 1 / 20 / 7 / 12
Woodpigeon / 4 / 11 / 20 / 25 / 10 / 2 / 6 / 14 / 13
House sparrow / 12 / 14 / 5 / 25 / 6 / 6 / 13 / 12
Yellowhammer / 12 / 14 / 4 / 17 / 13 / 2 / 13 / 1 / 14
Stonechat / 13 / 10 / 6 / 22 / 7 / 7 / 22
Dunnock / 4 / 13 / 9 / 21 / 5 / 11 / 4 / 11
Pheasant / 9 / 6 / 10 / 13 / 12 / 5 / 5 / 8 / 7
Table L1: (continued).
Chaffinch / 5 / 9 / 7 / 21 / 9 / 11 / 1
Greenfinch / 1 / 14 / 15 / 15
Whitethroat / 9 / 5 / 1 / 7 / 8 / 2 / 6 / 7
Spotted flycatcher / 12 / 1 / 13 / 6 / 7
Whinchat / 11 / 2 / 5 / 8 / 8 / 5
Cirl bunting / 8 / 2 / 2 / 9 / 3 / 1 / 2 / 9
Pied wagtail / 9 / 1 / 8 / 2 / 3 / 5 / 2
Great tit / 4 / 4 / 4
Chiffchaff / 1 / 2 / 2 / 1 / 3
Mistle thrush / 2 / 1 / 3 / 3
Raven / 3 / 3 / 3
Wren / 2 / 1 / 1 / 2 / 2 / 1
Buzzard / 1 / 1 / 1 / 1 / 1 / 1
Great spotted woodpecker / 2 / 2 / 2
Jay / 1 / 1 / 2 / 1 / 1
Song thrush / 1 / 1 / 1 / 1 / 2
Green woodpecker / 1 / 1 / 1
Red-legged partridge / 1 / 1 / 1
Sand martin / 1 / 1 / 1
Willow warbler / 1 / 1 / 1
Table L2: Bird usage on Experiment 1 sites during May-September. Cumulative totals of foraging visits by each guild are tabulated for each treatment, closure period (pre- and post-closure) and year. Guilds are sorted in descending order of overall usage.
grouping / species/guild / Lenient / Moderate / Control / PRE / POST / 2006 / 2007 / 2008 / 2009guild / aerial feeders / 1182 / 1370 / 1479 / 1599 / 2432 / 668 / 1542 / 1034 / 787
obligate granivores / 96 / 257 / 372 / 168 / 557 / 216 / 328 / 85 / 96
small insectivores / 269 / 238 / 202 / 287 / 422 / 141 / 244 / 198 / 126
thrushes and starling / 110 / 118 / 126 / 277 / 77 / 7 / 63 / 147 / 137
corvids / 55 / 185 / 100 / 132 / 208 / 22 / 192 / 75 / 51
mixed diet granivores (excluding skylark) / 37 / 39 / 18 / 72 / 22 / 2 / 29 / 27 / 36
buntings / 20 / 16 / 6 / 26 / 16 / 2 / 14 / 3 / 23
pigeons & doves / 4 / 11 / 20 / 25 / 10 / 2 / 6 / 14 / 13
gamebirds / 9 / 6 / 11 / 13 / 13 / 5 / 5 / 9 / 7
others / 2 / 1 / 2 / 1 / 1 / 2
raptors / 1 / 1 / 1 / 1 / 1 / 1
Table L3: Bird usage on Experiment 2 - Devon Extension sites during May-September. Cumulative totals of foraging visits by each species are tabulated for each treatment, closure period (pre- and post-closure) and year. Species are sorted in descending order of overall usage.
grouping / species/guild / Lenient / Control / PRE / POST / 2008 / 2009species / Swallow / 480 / 617 / 848 / 249 / 832 / 265
House martin / 102 / 210 / 161 / 151 / 299 / 13
Skylark / 162 / 50 / 123 / 89 / 62 / 150
Blackbird / 44 / 53 / 89 / 8 / 70 / 27
Woodpigeon / 29 / 67 / 82 / 14 / 41 / 55
Carrion crow / 16 / 57 / 64 / 9 / 21 / 52
Robin / 35 / 36 / 46 / 25 / 49 / 22
Goldfinch / 35 / 9 / 19 / 25 / 28 / 16
Linnet / 25 / 12 / 33 / 4 / 20 / 17
Yellowhammer / 12 / 21 / 18 / 15 / 3 / 30
House sparrow / 12 / 11 / 19 / 4 / 6 / 17
Swift / 8 / 9 / 13 / 4 / 4 / 13
Dunnock / 2 / 11 / 13 / 3 / 10
Magpie / 12 / 12 / 2 / 10
Cirl bunting / 9 / 2 / 10 / 1 / 6 / 5
Starling / 8 / 8 / 8
Greenfinch / 1 / 5 / 4 / 2 / 3 / 3
Whitethroat / 4 / 2 / 4 / 2 / 1 / 5
Pied wagtail / 6 / 6 / 6
Mistle thrush / 5 / 5 / 5
Pheasant / 2 / 1 / 3 / 3
Chaffinch / 1 / 2 / 3 / 3
Bullfinch / 3 / 3 / 3
Herring gull / 3 / 3 / 3
Table L3: continued.
grouping / species/guild / Lenient / Control / PRE / POST / 2008 / 2009Jackdaw / 2 / 2 / 2
Wren / 1 / 1 / 2 / 1 / 1
Meadow pipit / 1 / 1 / 1
Stonechat / 1 / 1 / 1
Buzzard / 1 / 1 / 1
Stock dove / 1 / 1 / 1
Table L4: Bird usage on Experiment 2 - Devon Extension sites during May-September. Cumulative totals of foraging visits by each guild are tabulated for each treatment, closure period (pre- and post-closure) and year. Guilds are sorted in descending order of overall usage.
grouping / species/guild / Lenient / Control / PRE / POST / 2008 / 2009guild / aerial feeders / 590 / 836 / 1022 / 404 / 1135 / 291
thrushes and starling / 44 / 66 / 102 / 8 / 75 / 35
small insectivores / 43 / 57 / 70 / 30 / 60 / 40
pigeons & doves / 29 / 68 / 83 / 14 / 42 / 55
corvids / 16 / 71 / 78 / 9 / 25 / 62
obligate granivores / 61 / 26 / 56 / 31 / 51 / 36
mixed diet granivores (excluding skylark) / 37 / 36 / 50 / 23 / 18 / 55
buntings / 21 / 23 / 28 / 16 / 9 / 35
gamebirds / 2 / 1 / 3 / 3
others / 3 / 3 / 3
raptors / 1 / 1 / 1
Table L5: Bird usage on Experiment 1 sites during winter 2007/08. For each species, cumulative totals of foraging visits and the number of plot-visits with non-zero counts are tabulated for each treatment. Species are sorted in descending order of overall usage.
grouping / Species/guild / Lenient / Moderate / Control / Lenient / Moderate / Control
species / Meadow pipit / 204 / 273 / 77 / 19 / 18 / 8
Yellowhammer / 17 / 24 / 0 / 2 / 1
Robin / 2 / 20 / 16 / 2 / 2 / 2
Blackbird / 6 / 4 / 13 / 4 / 2 / 3
Snipe / 4 / 6 / 4 / 2 / 2 / 2
Stonechat / 1 / 12 / 0 / 1 / 2
Woodpigeon / 0 / 0 / 10 / 2
Cirl bunting / 6 / 0 / 0 / 1
Song thrush / 1 / 1 / 4 / 1 / 1 / 3
Carrion crow / 3 / 1 / 1 / 2 / 1 / 1
Magpie / 2 / 2 / 1 / 1 / 1 / 1
Rook / 0 / 0 / 5 / 1
Great tit / 0 / 3 / 0 / 1
Pheasant / 1 / 0 / 1 / 1 / 1
Wren / 0 / 2 / 0 / 1
Bullfinch / 0 / 0 / 1 / 1
Skylark / 1 / 0 / 0 / 1
Table L6: Bird usage on Experiment 1 sites during winter 2007/08. For each guild, cumulative totals of foraging visits and the number of plot-visits with non-zero counts are tabulated for each treatment. Guilds are sorted in descending order of overall usage.
cumulative counts / non-zero countsgrouping / Species/guild / Lenient / Moderate / Control / Lenient / Moderate / Control
guild / small insectivores / 207 / 310 / 93 / 19 / 19 / 9
mixed diet granivores / 24 / 24 / 1 / 4 / 1 / 1
thrushes and starling / 7 / 5 / 17 / 4 / 3 / 4
corvids / 5 / 3 / 7 / 2 / 1 / 3
others / 4 / 6 / 4 / 2 / 2 / 2
pigeons & doves / 0 / 0 / 10 / 2
gamebirds / 1 / 0 / 1 / 1 / 1
10
Appendix M: Bird usage models
1. Breeding season treatment models (Experiment 1 and Devon Extension)
1.1 Initial analysis of treatment responses
Treatment responses across the species/guilds were compared initially by modelling the number of foraging visits per timed watch in negative binomial errors GLMs containing site as a fixed effect (Tables M1.1 & M1.2). Sites where the species or guild did not occur were excluded from the models as these data contained no information on treatment responses. Treatment*closure period interactions were tested and were only significant for skylarks on Experiment 1 sites.
Table M1.1: Treatment models for the main species and guilds occurring on Experiment 1 sites during the breeding season. Probability values for tests of the main treatment effect and post hoc tests (least squares) between the treatments are presented, along with least squares estimates of mean usage of each treatment (foraging visits per timed watch).
Means / Post hoc contrastsSpecies/guild / No.farms / P(treatment) / Lenient / Moderate / Control / C-L / C-M / L-M
Aerial feeders / 13 / 0.21 / 2.75 / 3.58 / 3.73 / 0.06 / 0.79 / 0.10
Buntings / 6 / 0.14 / 0.07 / 0.17 / 0.07 / 0.05 / 0.15 / 0.56
Corvids / 12 / 0.002 / 0.14 / 0.52 / 0.31 / 0.04 / 0.15 / 0.00
Mixed diet granivores (excluding Skylark) / 10 / 0.10 / 0.26 / 0.23 / 0.10 / 0.05 / 0.07 / 0.82
Obligate granivores / 13 / 0.64 / 0.26 / 0.34 / 0.39 / 0.36 / 0.75 / 0.51
Small insectivores / 13 / 0.29 / 0.73 / 0.76 / 0.52 / 0.20 / 0.15 / 0.88
Blackbird / 12 / 0.90 / 0.39 / 0.41 / 0.35 / 0.77 / 0.64 / 0.86
Meadow pipit / 6 / 0.20 / 0.65 / 0.29 / 0.15 / 0.07 / 0.39 / 0.31
Robin / 12 / 0.25 / 0.40 / 0.50 / 0.29 / 0.33 / 0.10 / 0.48
Starling / 6 / 0.66 / 0.46 / 0.19 / 0.11 / 0.36 / 0.70 / 0.54
Skylark / 8 / <0.0001 / 0.99 / 0.34 / 0.15 / <0.0001 / 0.00 / <0.0001
Woodpigeon / 7 / 0.05 / 0.04 / 0.10 / 0.22 / 0.01 / 0.17 / 0.20
Table M1.2: Treatment models for the main species and guilds occurring on Experiment 2 - Devon Extension sites during the breeding season. Tests of the main treatment effect and least squares estimates of mean usage of each treatment (foraging visits per timed watch) are tabulated.
Species/guild / No.farms / P(treatment) / Lenient / Control
Aerial feeders / 8 / 0.19 / 4.46 / 6.06
Buntings / 4 / 0.34 / 0.32 / 0.18
Corvids / 6 / 0.0003 / 0.08 / 0.36
Mixed diet granivores (excluding Skylark) / 5 / 0.56 / 0.38 / 0.29
Obligate granivores / 6 / 0.07 / 0.86 / 0.36
Small insectivores / 8 / 0.39 / 0.31 / 0.46
Blackbird / 7 / 0.75 / 0.30 / 0.34
Skylark / 6 / <0.0001 / 0.87 / 0.24
1.2 Full treatment models for priority bird groups
The priority species that showed treatment responses in this initial analysis were examined in greater detail. Data for skylarks, buntings and obligate granivores were modelled using negative binomial errors GLMMs with site as a random effect or GLMs with site as a fixed effect (where birds occurred at too few sites to fit random site effects). These models were used to test for interactions between treatment and closure period (before/after closure), year or site. The resulting models are summarised in the main report. The guild of mixed diet granivores (excluding skylarks) incorporated additional species with foraging requirements that were similar to buntings in an attempt to improve model precision. This approach did not improve upon the buntings-only model, so it is not reported further.