MyczkoŁ, RosinZM, SkórkaP, WylegałaP, TobolkaM, FliszkiewiczM, MizeraT, Tryjanowski P
Effects of management intensity on orchard features and bird communities in winter
SUPPLEMENTARY FILE
1. Correlation coefficients between continuous explanatory variables in orchards differing in management intensity. Statistically significant correlations are in bold.
Correlation coefficients between continuous environmental variables in abandoned orchards (n = 38). For explanation of variable codes: see Table 1 in main text.
Area(1) / NotMown
(2) / OrchCov
(3) / GrassCov
(4) / BuildCov
(5) / ForCov
(6) / FruitDen
(7) / TreeDen
(8) / TreeDiv
(9) / LandDiv
(11)
(1) / - / -0.026 (0.893) / 0.029 (0.863) / -0.104 (0.534) / -0.150 (0.369) / 0.101
(0.546) / 0.064 (0.703) / -0.181 (0.277) / -0.304 (0.064) / -0.024 (0.887)
(2) / - / -0.194 (0.242) / 0.155 (0.354) / 0.162 (0.331) / 0.167 (0.315) / 0.050 (0.764) / -0.276 (0.093) / 0.038 (0.819) / 0.065 (0.699)
(3) / - / -0.077 (0.647) / 0.016 (0.923) / 0.027 (0.874) / 0.064 (0.702) / 0.309 (0.059) / -0.122 (0.465) / 0.372 (0.021)
(4) / - / -0.184 (0.269) / 0.204 (0.218) / 0.226 (0.173) / 0.062 (0.710) / 0.356 (0.028) / 0.139 (0.407)
(5) / - / -0.232 (0.161) / 0.061 (0.715) / -0.006 (0.971) / 0.113 (0.499) / 0.366 (0.024)
(6) / - / -0.193 (0.245) / -0.446 (0.005) / 0.107 (0.524) / 0.494 (0.002)
(7) / - / 0.304 (0.063) / 0.311 (0.058) / -0.026 (0.875)
(8) / - / 0.079
(0.637) / -0.199 (0.231)
(9) / - / 0.093 (0.577)
(10) / -
Correlation coefficients between continuous environmental variables in extensive orchards (n = 45). For explanation of variable codes: see Table 1 in main text.
Area(1) / NotMown
(2) / OrchCov
(3) / GrassCov
(4) / BuildCov
(5) / ForCov
(6) / FruitDen
(7) / TreeDen
(8) / TreeDiv
(9) / LandDiv
(11)
(1) / - / 0.390 (0.008) / 0.424 (0.004) / -0.163 (0.284) / -0.078 (0.609) / -0.063 (0.680) / -0.017 (0.914) / -0.109 (0.475) / -0.186 (0.221) / 0.169
(0.268)
(2) / - / 0.302 (0.044) / 0.322 (0.031) / -0.144 (0.345) / 0.215 (0.157) / 0.128 (0.403) / 0.216 (0.155) / 0.078 (0.611) / 0.139 (0.361)
(3) / - / 0.132 (0.388) / -0.116 (0.447) / -0.053 (0.730) / 0.026 (0.867) / 0.275 (0.067) / -0.015 (0.922) / 0.435 (0.003)
(4) / - / -0.313 (0.037) / 0.206 (0.174) / 0.046 (0.763) / 0.063 (0.680) / -0.058 (0.707) / 0.412 (0.004)
(5) / - / -0.031 (0.843) / 0.240 (0.112) / 0.073 (0.631) / 0.026 (0.867) / 0.233 (0.122)
(6) / - / -0.181 (0.235) / 0.566 (<0.001) / 0.263 (0.081) / 0.140 (0.360)
(7) / - / -0.127 (0.406) / -0.011 (0.943) / 0.032 (0.836)
(8) / - / 0.508 (<0.001) / -0.125 (0.415)
(9) / - / -0.266 (0.078)
(10) / -
Correlation coefficients between continuous environmental variables in intensive orchards (n = 23). For explanation of variable codes: see Table 1 in main text.
Area(1) / NotMown
(2) / OrchCov
(3) / GrassCov
(4) / BuildCov
(5) / ForCov
(6) / FruitDen
(7) / TreeDen
(8) / TreeDiv
(9) / LandDiv
(11)
(1) / - / -0.450 (0.031) / 0.150 (0.493) / -0.124 (0.572) / -0.703 (<0.001) / 0.184 (0.349) / 0.087 (0.694) / 0.428 (0.042) / 0.339 (0.113) / -0.198 (0.366)
(2) / - / -0.124 (0.584) / 0.531 (0.009) / 0.348 (0.103) / 0.349 (0.102) / -0.023 (0.916) / -0.575 (0.004) / 0.007 (0.974) / 0.387 (0.068)
(3) / - / 0.296 (0.170) / -0.163 (0.459) / -0.075 (0.733) / -0.145 (0.510) / 0.033 (0.881) / 0.286 (0.186) / 0.463 (0.026)
(4) / - / 0.230 (0.291) / 0.336
(0.117) / -0.109 (0.621) / -0.378 (0.075) / 0.073 (0.739) / 0.770 (<0.001)
(5) / - / -0.268 (0.216) / -0.167 (0.446) / -0.374 (0.078) / -0.419 (0.046) / 0.243 (0.263)
(6) / - / 0.397 (0.061) / -0.042 (0.848) / 0.258 (0.235) / 0.527 (0.010)
(7) / - / 0.300 (0.164) / -0.277 (0.201) / 0.031 (0.887)
(8) / - / 0.035 (0.874) / -0.241 (0.267)
(9) / - / 0.164 (0.456)
(10) / -
Correlation coefficients between continuous environmental variables in all orchards (n = 106). For explanation of variable codes: see Table 1 in main text.
Area(1) / NotMown
(2) / OrchCov
(3) / GrassCov
(4) / BuildCov
(5) / ForCov
(6) / FruitDen
(7) / TreeDen
(8) / TreeDiv
(9) / LandDiv
(11)
(1) / - / -0.111 (0.256) / 0.399 (<0.001) / -0.010 (0.917) / -0.381 (<0.001) / 0.021 (0.829) / -0.103 (0.291) / -0.014 (0.888) / -0.291 (0.002) / 0.094 (0.340)
(2) / - / -0.091 (0.350) / 0.217 (0.026) / 0.102 (0.296) / 0.190 (0.051) / 0.124 (0.204) / 0.029 (0.767) / 0.132 (0.178) / 0.078 (0.425)
(3) / - / 0.162 (0.096) / -0.193 (0.047) / -0.002 (0.987) / -0.134 (0.172) / 0.204 (0.036) / -0.137 (0.162) / 0.453 (<0.001)
(4) / - / -0.177 (0.070) / 0.220 (0.024) / 0.024 (0.809) / 0.018 (0.852) / 0.037 (0.703) / 0.438 (<0.001)
(5) / - / -0.137 (0.163) / 0.143 (0.144) / 0.008 (0.938) / 0.121 (0.216) / 0.209 (0.032)
(6) / - / -0.061 (0.537) / 0.371 (<0.001) / 0.204 (0.036) / 0.326 (0.001)
(7) / - / -0.042 (0.671) / 0.135 (0.168) / -0.028 (0.772)
(8) / - / 0.455 (<0.001) / -0.085
(0.385)
(9) / - / -0.125 (0.202)
(10) / -
2. Alternative analysis on factors affecting species richness and density of birds in orchards.
Factors affecting bird species richness. Results of the generalized linear model with Poisson error variance. Significant effects are in bold. For explanation of variable codes see Table 1 in main text.
OrchType / 0.957 / 2 / 0.620
Area / 18.458 / 1 / <0.001
NotMown / 7.814 / 1 / 0.005
FruitDens / 0.978 / 1 / 0.323
TreeDen / 4.846 / 1 / 0.028
TreeDiv / 7.125 / 1 / 0.008
LandDiv / 16.972 / 1 / <0.001
OrchType * Area / 5.667 / 2 / 0.059
OrchType * NotMown / 0.059 / 2 / 0.971
OrchType * FruitDens / 0.010 / 2 / 0.995
OrchType * TreeDen / 9.692 / 2 / 0.008
OrchType * TreeDiv / 1.315 / 2 / 0.518
OrchType * LandDiv / 6.010 / 2 / 0.050
Factors affecting bird densities. Results of the generalized linear model with negative binomial error variance. Significant effects are in bold.For explanation of variable codes see Table 1 in main text.
Variable / χ2 / df / pOrchType / 6.991 / 2 / 0.030
Area / 6.440 / 1 / 0.011
NotMown / 7.986 / 1 / 0.005
FruitDens / 5.215 / 1 / 0.022
TreeDen / 1.349 / 1 / 0.245
TreeDiv / 5.180 / 1 / 0.023
LandDiv / 3.292 / 1 / 0.070
OrchType * Area / 0.999 / 2 / 0.607
OrchType * NotMown / 0.561 / 2 / 0.755
OrchType * FruitDens / 13.025 / 2 / <0.001
OrchType * TreeDen / 8.526 / 2 / 0.014
OrchType * TreeDiv / 3.579 / 2 / 0.167
OrchType * LandDiv / 18.370 / 2 / <0.001
3. Analysis on the estimated number of species.
We calculated number of species with different methods implemented in the software Spade (available from: and bias-corrected Chao 1 estimate (Shen et al. 2003, Chao 2005, Chao et al. 2006) was chosen (other methods did not calculate number of species from small samples). The estimated and observed numbers of species were strongly positively correlated (see graph below). The results of the general linear model with factors affecting estimated species number are also given below. The results, as expected form the correlation, were similar to that from the analysis on raw data.
All methods of accounting the imperfect detection have specific assumptions (see: Gotelli and Colwell 2010). Number of species recorded per orchard was very low, and in most cases below 10 which is a rule-of-thumb value below which calculations usually give imprecise and biased estimates. The assumption on random and independent distribution of individuals in orchards is obviously violated. Birds in winter often gather in intra- and interspecific flocks which make specimens both individually and spatially dependent. When, the assumptions of the methods are not met the estimates of number of species may be highly biased or/and imprecise. Thus, results presented below should be treated with caution because they may be not valid.
Variables affecting estimated species richness (result from the stepwise forward selection based on AICc).
Variable code / Estimated species richnessdf / F / p
OrchType / 2 / 4.218 / 0.043
Area / 1 / 11.929 / <0.001
NotMown / 1 / 5.952 / 0.016
FruitDen / 1 / 2.480 / 0.118
TreeDen / 1 / 2.376 / 0.126
TreeDiv / 1 / 4.263 / 0.042
LandDiv / 1 / 6.061 / 0.016
LandDiv*OrchType / 2 / 3.4575 / 0.066
References
Chao A (2005) Species estimation and applications. Encyclopedia of Statistical Sciences, Second Edition, Vol. 12, 7907-7916 (Balakrishnan N, Read CB, Vidakovic B, Editors),Wiley, New York.
Chao A, Shen T-J, Hwang WH (2006) Application of Laplace’s boundary-mode approximations to estimate species and shared species richness. Australian and New Zealand Journal of Statistics 48:117-128.
Gotelli NJ, Colwell RK (2010) Estimating species richness. In: Magurran AE, Mcgill BJ ed. Frontiers in Measuring Biodiversity. Oxford University Press, New York, pp 39–54.
Shen T-J, Chao A, Lin J-F (2002) Predicting the number of new species in further taxonomic sampling. Ecology 84:798-804