Analisis de comunidades- Vegan
Julian Avila-Jimenez
28/9/2020
Calculo de indice de disimilaridad
Existen diversos indices de disimilaridad para el analisis de comunidades basados en distancias estadisticas multidimensionales que permiten hacer comparaciones entre comunidades en funcion de las especies que las componen. En este ejercicio vamos a comparar diferentes sitios de estudio en funcion de la abundancia de las especies encontradas, primero hallamos un indice de disimilaridad de las comunidades Bray–Curtis y luego realizaremos un analisis de similitud anosim tomando la matriz de disimilitud y una variable discreta del set de datos ambiental.
Usando el tipo de manejo tenemos:
dune.dist<- vegdist(dune) #calculo de matriz de distancias. por defecto Bray–Curtis
dune.ano<- anosim(dune.dist, dune.env$Management)
summary(dune.ano)##
## Call:
## anosim(x = dune.dist, grouping = dune.env$Management)
## Dissimilarity: bray
##
## ANOSIM statistic R: 0.2579
## Significance: 0.013
##
## Permutation: free
## Number of permutations: 999
##
## Upper quantiles of permutations (null model):
## 90% 95% 97.5% 99%
## 0.121 0.169 0.218 0.264
##
## Dissimilarity ranks between and within classes:
## 0% 25% 50% 75% 100% N
## Between 4 58.50 104.00 145.500 188.0 147
## BF 5 15.25 25.50 41.250 57.0 3
## HF 1 7.25 46.25 68.125 89.5 10
## NM 6 64.75 124.50 156.250 181.0 15
## SF 3 32.75 53.50 99.250 184.0 15plot(dune.ano)## Warning in bxp(list(stats = structure(c(4, 58.5, 104, 145.5, 188, 5, 15.25, :
## some notches went outside hinges ('box'): maybe set notch=FALSE
El valor de la significancia nos muestra que existen comunidades que son diferentes entre si (no debe interpretarse como que todas las comunidades son diferentes).
Notese en el grafico que la comparacion debe realizarse entre cada uno de los grupos (Management) vs. todos los grupos (conjunto de todas las distancias), las cunas representan intervalos de confianza de 95% y cuando estas se sobreponen con las del grupo Between podemos inferir que no hay diferencia entre los grupos. En este caso encontramos diferencias en los grupos BF, HF y SF.
Analisis de porcentaje de similaridad
Este analisis nos permite ver cuales son las especies que generan diferencias entre los grupos y cual es el aporte de cada una de estas a la diferencia, el analisis se realiza en forma de contrastes entre grupos, asi podemos ver que tan diferente es un grupo particular de otro.
sim<- simper(dune, dune.env$Management)
summary(sim) ##
## Contrast: SF_BF
##
## average sd ratio ava avb cumsum
## Agrostol 0.061374 0.034193 1.7949 4.6667 0.0000 0.09824
## Alopgeni 0.052667 0.036476 1.4439 4.3333 0.6667 0.18255
## Lolipere 0.048116 0.039445 1.2198 3.0000 6.0000 0.25957
## Trifrepe 0.046297 0.025525 1.8138 1.3333 4.6667 0.33368
## Poatriv 0.046020 0.033801 1.3615 4.6667 3.6667 0.40734
## Scorautu 0.043697 0.024922 1.7534 1.3333 4.3333 0.47729
## Bromhord 0.033677 0.025860 1.3023 0.5000 2.6667 0.53120
## Achimill 0.030152 0.020821 1.4482 0.1667 2.3333 0.57947
## Planlanc 0.028585 0.021549 1.3265 0.0000 2.0000 0.62522
## Elymrepe 0.028074 0.029778 0.9428 2.0000 1.3333 0.67016
## Bracruta 0.025501 0.023902 1.0669 2.0000 2.0000 0.71098
## Poaprat 0.025129 0.023967 1.0485 2.5000 4.0000 0.75121
## Sagiproc 0.024326 0.022149 1.0983 1.8333 0.6667 0.79014
## Bellpere 0.019859 0.017088 1.1622 0.6667 1.6667 0.82193
## Eleopalu 0.018611 0.042958 0.4333 1.3333 0.0000 0.85172
## Anthodor 0.017543 0.025804 0.6798 0.0000 1.3333 0.87981
## Juncbufo 0.016031 0.023708 0.6762 1.1667 0.0000 0.90547
## Vicilath 0.014671 0.013306 1.1026 0.0000 1.0000 0.92895
## Hyporadi 0.010286 0.015198 0.6768 0.0000 0.6667 0.94542
## Ranuflam 0.009306 0.013595 0.6845 0.6667 0.0000 0.96031
## Juncarti 0.006979 0.016109 0.4333 0.5000 0.0000 0.97148
## Callcusp 0.006979 0.016109 0.4333 0.5000 0.0000 0.98266
## Rumeacet 0.004526 0.010444 0.4333 0.3333 0.0000 0.98990
## Cirsarve 0.003983 0.009185 0.4336 0.3333 0.0000 0.99628
## Chenalbu 0.002326 0.005370 0.4333 0.1667 0.0000 1.00000
## Airaprae 0.000000 0.000000 NaN 0.0000 0.0000 1.00000
## Comapalu 0.000000 0.000000 NaN 0.0000 0.0000 1.00000
## Empenigr 0.000000 0.000000 NaN 0.0000 0.0000 1.00000
## Salirepe 0.000000 0.000000 NaN 0.0000 0.0000 1.00000
## Trifprat 0.000000 0.000000 NaN 0.0000 0.0000 1.00000
##
## Contrast: SF_HF
##
## average sd ratio ava avb cumsum
## Agrostol 0.047380 0.031273 1.5151 4.6667 1.4 0.08351
## Alopgeni 0.046433 0.032897 1.4115 4.3333 1.6 0.16535
## Lolipere 0.041986 0.027007 1.5546 3.0000 4.0 0.23935
## Planlanc 0.039198 0.033208 1.1804 0.0000 3.0 0.30844
## Rumeacet 0.038992 0.027369 1.4247 0.3333 3.2 0.37716
## Elymrepe 0.031877 0.029550 1.0787 2.0000 2.0 0.43334
## Poatriv 0.028466 0.021522 1.3227 4.6667 4.8 0.48352
## Bracruta 0.025261 0.021044 1.2004 2.0000 2.8 0.52804
## Eleopalu 0.024974 0.038877 0.6424 1.3333 0.8 0.57206
## Poaprat 0.023932 0.019180 1.2478 2.5000 3.4 0.61424
## Anthodor 0.023409 0.021430 1.0923 0.0000 1.8 0.65550
## Sagiproc 0.023144 0.020479 1.1301 1.8333 0.8 0.69629
## Trifprat 0.023080 0.023432 0.9850 0.0000 1.8 0.73697
## Juncarti 0.022850 0.025677 0.8899 0.5000 1.6 0.77724
## Trifrepe 0.022383 0.019487 1.1486 1.3333 2.8 0.81669
## Juncbufo 0.021643 0.022237 0.9733 1.1667 1.2 0.85484
## Scorautu 0.020509 0.016422 1.2489 1.3333 2.8 0.89099
## Achimill 0.015183 0.011393 1.3326 0.1667 1.2 0.91775
## Bromhord 0.013375 0.014504 0.9222 0.5000 0.8 0.94132
## Ranuflam 0.010661 0.013387 0.7964 0.6667 0.4 0.96011
## Bellpere 0.009991 0.012571 0.7948 0.6667 0.4 0.97772
## Callcusp 0.006623 0.015076 0.4393 0.5000 0.0 0.98939
## Cirsarve 0.003809 0.008669 0.4394 0.3333 0.0 0.99611
## Chenalbu 0.002208 0.005025 0.4393 0.1667 0.0 1.00000
## Airaprae 0.000000 0.000000 NaN 0.0000 0.0 1.00000
## Comapalu 0.000000 0.000000 NaN 0.0000 0.0 1.00000
## Empenigr 0.000000 0.000000 NaN 0.0000 0.0 1.00000
## Hyporadi 0.000000 0.000000 NaN 0.0000 0.0 1.00000
## Salirepe 0.000000 0.000000 NaN 0.0000 0.0 1.00000
## Vicilath 0.000000 0.000000 NaN 0.0000 0.0 1.00000
##
## Contrast: SF_NM
##
## average sd ratio ava avb cumsum
## Poatriv 0.078284 0.040947 1.9118 4.6667 0.0000 0.1014
## Alopgeni 0.071219 0.046958 1.5167 4.3333 0.0000 0.1936
## Agrostol 0.056508 0.044176 1.2792 4.6667 2.1667 0.2667
## Lolipere 0.054851 0.059914 0.9155 3.0000 0.3333 0.3378
## Eleopalu 0.048027 0.047168 1.0182 1.3333 2.1667 0.3999
## Poaprat 0.040724 0.031790 1.2810 2.5000 0.6667 0.4527
## Bracruta 0.040008 0.034398 1.1631 2.0000 2.8333 0.5045
## Elymrepe 0.035598 0.038515 0.9243 2.0000 0.0000 0.5506
## Scorautu 0.032475 0.034813 0.9328 1.3333 3.1667 0.5926
## Trifrepe 0.030430 0.031634 0.9619 1.3333 1.8333 0.6320
## Sagiproc 0.030304 0.030477 0.9943 1.8333 0.5000 0.6712
## Salirepe 0.029275 0.032014 0.9144 0.0000 1.8333 0.7092
## Anthodor 0.024541 0.036694 0.6688 0.0000 1.3333 0.7409
## Callcusp 0.022763 0.029443 0.7731 0.5000 1.1667 0.7704
## Ranuflam 0.022566 0.022819 0.9889 0.6667 1.3333 0.7996
## Juncarti 0.022543 0.028598 0.7883 0.5000 1.1667 0.8288
## Hyporadi 0.020108 0.031291 0.6426 0.0000 1.1667 0.8548
## Juncbufo 0.019860 0.029034 0.6840 1.1667 0.0000 0.8806
## Planlanc 0.015420 0.022769 0.6772 0.0000 0.8333 0.9005
## Airaprae 0.014883 0.021881 0.6802 0.0000 0.8333 0.9198
## Bellpere 0.012317 0.015921 0.7737 0.6667 0.3333 0.9357
## Comapalu 0.011883 0.017407 0.6826 0.0000 0.6667 0.9511
## Achimill 0.009294 0.014931 0.6224 0.1667 0.3333 0.9632
## Bromhord 0.007172 0.016333 0.4391 0.5000 0.0000 0.9724
## Rumeacet 0.005590 0.012751 0.4384 0.3333 0.0000 0.9797
## Empenigr 0.005225 0.012001 0.4354 0.0000 0.3333 0.9864
## Cirsarve 0.004782 0.010889 0.4391 0.3333 0.0000 0.9926
## Chenalbu 0.002893 0.006602 0.4382 0.1667 0.0000 0.9964
## Vicilath 0.002792 0.006425 0.4345 0.0000 0.1667 1.0000
## Trifprat 0.000000 0.000000 NaN 0.0000 0.0000 1.0000
##
## Contrast: BF_HF
##
## average sd ratio ava avb cumsum
## Rumeacet 0.038666 0.02606 1.4838 0.0000 3.2 0.08163
## Poatriv 0.033301 0.02579 1.2911 3.6667 4.8 0.15194
## Planlanc 0.031852 0.01830 1.7401 2.0000 3.0 0.21918
## Bromhord 0.028651 0.01799 1.5926 2.6667 0.8 0.27967
## Lolipere 0.028431 0.02215 1.2834 6.0000 4.0 0.33970
## Elymrepe 0.027822 0.02959 0.9404 1.3333 2.0 0.39843
## Trifrepe 0.025838 0.01656 1.5603 4.6667 2.8 0.45298
## Anthodor 0.023582 0.02042 1.1547 1.3333 1.8 0.50277
## Achimill 0.023426 0.01474 1.5893 2.3333 1.2 0.55223
## Bracruta 0.022733 0.01802 1.2617 2.0000 2.8 0.60022
## Alopgeni 0.021610 0.02308 0.9363 0.6667 1.6 0.64584
## Trifprat 0.021514 0.02207 0.9747 0.0000 1.8 0.69126
## Juncarti 0.020084 0.02555 0.7860 0.0000 1.6 0.73367
## Scorautu 0.019318 0.01356 1.4241 4.3333 2.8 0.77445
## Bellpere 0.018290 0.01486 1.2305 1.6667 0.4 0.81306
## Agrostol 0.017605 0.02284 0.7708 0.0000 1.4 0.85023
## Juncbufo 0.015000 0.02066 0.7260 0.0000 1.2 0.88190
## Vicilath 0.012848 0.01140 1.1274 1.0000 0.0 0.90903
## Sagiproc 0.011685 0.01297 0.9008 0.6667 0.8 0.93369
## Eleopalu 0.010169 0.02111 0.4817 0.0000 0.8 0.95516
## Hyporadi 0.008950 0.01312 0.6824 0.6667 0.0 0.97406
## Poaprat 0.007203 0.01010 0.7133 4.0000 3.4 0.98927
## Ranuflam 0.005084 0.01055 0.4817 0.0000 0.4 1.00000
## Airaprae 0.000000 0.00000 NaN 0.0000 0.0 1.00000
## Chenalbu 0.000000 0.00000 NaN 0.0000 0.0 1.00000
## Cirsarve 0.000000 0.00000 NaN 0.0000 0.0 1.00000
## Comapalu 0.000000 0.00000 NaN 0.0000 0.0 1.00000
## Empenigr 0.000000 0.00000 NaN 0.0000 0.0 1.00000
## Salirepe 0.000000 0.00000 NaN 0.0000 0.0 1.00000
## Callcusp 0.000000 0.00000 NaN 0.0000 0.0 1.00000
##
## Contrast: BF_NM
##
## average sd ratio ava avb cumsum
## Lolipere 0.090681 0.02644 3.4292 6.0000 0.3333 0.1243
## Poatriv 0.054684 0.04465 1.2248 3.6667 0.0000 0.1992
## Poaprat 0.052511 0.01813 2.8966 4.0000 0.6667 0.2712
## Trifrepe 0.051287 0.02756 1.8611 4.6667 1.8333 0.3415
## Bromhord 0.039689 0.02920 1.3590 2.6667 0.0000 0.3959
## Bracruta 0.035723 0.02869 1.2452 2.0000 2.8333 0.4448
## Eleopalu 0.033759 0.03573 0.9449 0.0000 2.1667 0.4911
## Agrostol 0.033446 0.03473 0.9630 0.0000 2.1667 0.5369
## Achimill 0.033190 0.02338 1.4198 2.3333 0.3333 0.5824
## Scorautu 0.031356 0.02026 1.5480 4.3333 3.1667 0.6254
## Anthodor 0.028060 0.03295 0.8517 1.3333 1.3333 0.6638
## Planlanc 0.027319 0.02193 1.2458 2.0000 0.8333 0.7013
## Salirepe 0.026770 0.02927 0.9145 0.0000 1.8333 0.7379
## Bellpere 0.023529 0.01909 1.2322 1.6667 0.3333 0.7702
## Hyporadi 0.021721 0.02450 0.8864 0.6667 1.1667 0.8000
## Ranuflam 0.020314 0.02275 0.8928 0.0000 1.3333 0.8278
## Elymrepe 0.019993 0.02926 0.6833 1.3333 0.0000 0.8552
## Callcusp 0.017833 0.02681 0.6653 0.0000 1.1667 0.8796
## Juncarti 0.017694 0.02600 0.6806 0.0000 1.1667 0.9039
## Vicilath 0.015773 0.01447 1.0902 1.0000 0.1667 0.9255
## Sagiproc 0.015432 0.01857 0.8310 0.6667 0.5000 0.9466
## Airaprae 0.013410 0.01969 0.6809 0.0000 0.8333 0.9650
## Comapalu 0.010739 0.01571 0.6835 0.0000 0.6667 0.9797
## Alopgeni 0.009997 0.01463 0.6833 0.6667 0.0000 0.9934
## Empenigr 0.004787 0.01105 0.4332 0.0000 0.3333 1.0000
## Chenalbu 0.000000 0.00000 NaN 0.0000 0.0000 1.0000
## Cirsarve 0.000000 0.00000 NaN 0.0000 0.0000 1.0000
## Juncbufo 0.000000 0.00000 NaN 0.0000 0.0000 1.0000
## Rumeacet 0.000000 0.00000 NaN 0.0000 0.0000 1.0000
## Trifprat 0.000000 0.00000 NaN 0.0000 0.0000 1.0000
##
## Contrast: HF_NM
##
## average sd ratio ava avb cumsum
## Poatriv 0.071553 0.013681 5.2302 4.8 0.0000 0.09913
## Lolipere 0.054533 0.029625 1.8408 4.0 0.3333 0.17468
## Rumeacet 0.046546 0.030806 1.5109 3.2 0.0000 0.23917
## Poaprat 0.041750 0.018852 2.2146 3.4 0.6667 0.29701
## Planlanc 0.041633 0.029560 1.4084 3.0 0.8333 0.35469
## Bracruta 0.035340 0.020104 1.7579 2.8 2.8333 0.40365
## Eleopalu 0.032043 0.032315 0.9916 0.8 2.1667 0.44805
## Agrostol 0.031915 0.028887 1.1048 1.4 2.1667 0.49227
## Trifrepe 0.030372 0.022871 1.3280 2.8 1.8333 0.53434
## Elymrepe 0.029811 0.038676 0.7708 2.0 0.0000 0.57565
## Anthodor 0.028717 0.024799 1.1580 1.8 1.3333 0.61543
## Juncarti 0.027414 0.028537 0.9607 1.6 1.1667 0.65341
## Trifprat 0.025843 0.025972 0.9951 1.8 0.0000 0.68922
## Salirepe 0.025340 0.027291 0.9285 0.0 1.8333 0.72432
## Alopgeni 0.024459 0.032399 0.7549 1.6 0.0000 0.75821
## Scorautu 0.020703 0.014125 1.4658 2.8 3.1667 0.78689
## Ranuflam 0.019285 0.019939 0.9672 0.4 1.3333 0.81361
## Juncbufo 0.018181 0.024648 0.7376 1.2 0.0000 0.83880
## Hyporadi 0.017141 0.026548 0.6457 0.0 1.1667 0.86255
## Callcusp 0.016833 0.024901 0.6760 0.0 1.1667 0.88587
## Achimill 0.016555 0.014900 1.1111 1.2 0.3333 0.90881
## Sagiproc 0.015282 0.016535 0.9243 0.8 0.5000 0.92998
## Airaprae 0.012605 0.018243 0.6910 0.0 0.8333 0.94744
## Bromhord 0.012094 0.015169 0.7973 0.8 0.0000 0.96420
## Comapalu 0.010105 0.014556 0.6942 0.0 0.6667 0.97820
## Bellpere 0.008801 0.013732 0.6409 0.4 0.3333 0.99039
## Empenigr 0.004536 0.010325 0.4393 0.0 0.3333 0.99668
## Vicilath 0.002399 0.005461 0.4393 0.0 0.1667 1.00000
## Chenalbu 0.000000 0.000000 NaN 0.0 0.0000 1.00000
## Cirsarve 0.000000 0.000000 NaN 0.0 0.0000 1.00000
## Permutation: free
## Number of permutations: 0Usando el uso tenemos:
dune.ano<- anosim(dune.dist, dune.env$Use)
summary(dune.ano)##
## Call:
## anosim(x = dune.dist, grouping = dune.env$Use)
## Dissimilarity: bray
##
## ANOSIM statistic R: 0.05046
## Significance: 0.254
##
## Permutation: free
## Number of permutations: 999
##
## Upper quantiles of permutations (null model):
## 90% 95% 97.5% 99%
## 0.107 0.146 0.183 0.219
##
## Dissimilarity ranks between and within classes:
## 0% 25% 50% 75% 100% N
## Between 1.0 45.500 98.50 145.500 188 131
## Hayfield 15.0 68.000 107.00 134.000 181 21
## Haypastu 3.0 36.250 80.25 117.875 188 28
## Pasture 27.5 53.875 63.50 158.375 167 10plot(dune.ano)## Warning in bxp(list(stats = structure(c(1, 45.5, 98.5, 145.5, 188, 15, 68, :
## some notches went outside hinges ('box'): maybe set notch=FALSE
En este caso el valor de p y el grafico nos indica que en terminos de uso no hay diferencias entre los grupos.
sim<- simper(dune, dune.env$Use)
summary(sim)##
## Contrast: Haypastu_Hayfield
##
## average sd ratio ava avb cumsum
## Poatriv 0.055255 0.043137 1.2809 4.625 2.1429 0.08239
## Lolipere 0.050354 0.043090 1.1686 3.625 1.7143 0.15748
## Agrostol 0.044052 0.037897 1.1624 3.125 1.1429 0.22317
## Alopgeni 0.043356 0.044548 0.9732 3.000 0.4286 0.28782
## Anthodor 0.034739 0.032927 1.0550 0.375 2.2857 0.33962
## Bracruta 0.034551 0.030189 1.1445 2.250 2.7143 0.39114
## Elymrepe 0.034283 0.033794 1.0145 2.000 1.4286 0.44226
## Poaprat 0.030980 0.025271 1.2259 2.750 2.0000 0.48845
## Planlanc 0.029905 0.026591 1.1246 0.625 1.8571 0.53305
## Trifrepe 0.029817 0.023936 1.2457 2.375 2.1429 0.57751
## Scorautu 0.025736 0.028326 0.9086 2.250 3.2857 0.61588
## Salirepe 0.024668 0.030874 0.7990 0.000 1.5714 0.65267
## Sagiproc 0.023017 0.025613 0.8986 1.375 0.7143 0.68699
## Rumeacet 0.021550 0.028800 0.7483 1.000 1.0000 0.71912
## Juncarti 0.019593 0.027173 0.7211 0.375 1.1429 0.74834
## Achimill 0.019478 0.018795 1.0364 0.750 1.1429 0.77738
## Bromhord 0.018216 0.023106 0.7883 0.875 0.8571 0.80455
## Eleopalu 0.017683 0.033539 0.5272 0.625 0.5714 0.83091
## Juncbufo 0.017533 0.025943 0.6758 0.875 0.5714 0.85706
## Hyporadi 0.016928 0.029300 0.5777 0.000 1.0000 0.88230
## Bellpere 0.015810 0.016255 0.9726 0.875 0.8571 0.90588
## Ranuflam 0.014848 0.021865 0.6791 0.500 0.5714 0.92802
## Airaprae 0.012524 0.020556 0.6093 0.000 0.7143 0.94669
## Trifprat 0.010741 0.021097 0.5091 0.625 0.2857 0.96271
## Callcusp 0.006605 0.016585 0.3983 0.000 0.4286 0.97256
## Comapalu 0.004609 0.012582 0.3663 0.250 0.0000 0.97943
## Empenigr 0.004404 0.011057 0.3983 0.000 0.2857 0.98600
## Vicilath 0.004203 0.006896 0.6094 0.000 0.2857 0.99226
## Cirsarve 0.003253 0.008777 0.3707 0.250 0.0000 0.99711
## Chenalbu 0.001936 0.005247 0.3689 0.125 0.0000 1.00000
##
## Contrast: Haypastu_Pasture
##
## average sd ratio ava avb cumsum
## Poatriv 0.049381 0.038413 1.2855 4.625 2.2 0.08118
## Lolipere 0.048219 0.044834 1.0755 3.625 3.4 0.16045
## Eleopalu 0.047813 0.044108 1.0840 0.625 3.2 0.23905
## Alopgeni 0.044375 0.038944 1.1395 3.000 1.8 0.31199
## Agrostol 0.042701 0.036638 1.1655 3.125 3.0 0.38219
## Trifrepe 0.032276 0.030126 1.0714 2.375 2.6 0.43525
## Bracruta 0.030530 0.025732 1.1864 2.250 2.4 0.48544
## Poaprat 0.030333 0.029067 1.0436 2.750 2.4 0.53530
## Elymrepe 0.029633 0.031346 0.9453 2.000 0.0 0.58402
## Planlanc 0.025567 0.030339 0.8427 0.625 1.6 0.62604
## Scorautu 0.024920 0.022259 1.1195 2.250 2.6 0.66701
## Callcusp 0.022838 0.030162 0.7572 0.000 1.4 0.70455
## Sagiproc 0.022335 0.022654 0.9859 1.375 0.8 0.74127
## Juncarti 0.021381 0.025145 0.8503 0.375 1.4 0.77642
## Rumeacet 0.017941 0.025524 0.7029 1.000 0.6 0.80591
## Ranuflam 0.016398 0.015621 1.0498 0.500 1.2 0.83287
## Juncbufo 0.015681 0.021880 0.7167 0.875 0.4 0.85865
## Bromhord 0.014429 0.019715 0.7319 0.875 0.4 0.88237
## Achimill 0.012680 0.014748 0.8598 0.750 0.4 0.90321
## Trifprat 0.011945 0.020665 0.5780 0.625 0.4 0.92285
## Bellpere 0.011581 0.015730 0.7363 0.875 0.0 0.94189
## Comapalu 0.009262 0.015472 0.5986 0.250 0.4 0.95711
## Anthodor 0.008872 0.014402 0.6160 0.375 0.4 0.97170
## Hyporadi 0.006069 0.012495 0.4857 0.000 0.4 0.98167
## Vicilath 0.006069 0.012495 0.4857 0.000 0.4 0.99165
## Cirsarve 0.003191 0.008582 0.3719 0.250 0.0 0.99690
## Chenalbu 0.001887 0.005081 0.3714 0.125 0.0 1.00000
## Airaprae 0.000000 0.000000 NaN 0.000 0.0 1.00000
## Empenigr 0.000000 0.000000 NaN 0.000 0.0 1.00000
## Salirepe 0.000000 0.000000 NaN 0.000 0.0 1.00000
##
## Contrast: Hayfield_Pasture
##
## average sd ratio ava avb cumsum
## Eleopalu 0.048348 0.04723 1.0238 0.5714 3.2 0.07308
## Lolipere 0.048008 0.03889 1.2344 1.7143 3.4 0.14565
## Agrostol 0.045299 0.04084 1.1091 1.1429 3.0 0.21412
## Poatriv 0.036978 0.03121 1.1849 2.1429 2.2 0.27002
## Anthodor 0.034904 0.03083 1.1322 2.2857 0.4 0.32277
## Trifrepe 0.033616 0.03495 0.9618 2.1429 2.6 0.37359
## Planlanc 0.031371 0.02603 1.2054 1.8571 1.6 0.42101
## Poaprat 0.030990 0.02501 1.2393 2.0000 2.4 0.46785
## Bracruta 0.029668 0.02815 1.0538 2.7143 2.4 0.51270
## Scorautu 0.027753 0.02478 1.1199 3.2857 2.6 0.55465
## Alopgeni 0.027641 0.03201 0.8636 0.4286 1.8 0.59643
## Juncarti 0.025955 0.02803 0.9261 1.1429 1.4 0.63566
## Salirepe 0.024915 0.03088 0.8068 1.5714 0.0 0.67332
## Callcusp 0.024639 0.03020 0.8159 0.4286 1.4 0.71056
## Ranuflam 0.022428 0.01844 1.2161 0.5714 1.2 0.74446
## Hyporadi 0.019119 0.02716 0.7040 1.0000 0.4 0.77336
## Elymrepe 0.018868 0.03130 0.6028 1.4286 0.0 0.80188
## Rumeacet 0.018056 0.02431 0.7426 1.0000 0.6 0.82918
## Achimill 0.017924 0.02050 0.8744 1.1429 0.4 0.85627
## Sagiproc 0.015974 0.01737 0.9196 0.7143 0.8 0.88042
## Bromhord 0.014052 0.01913 0.7346 0.8571 0.4 0.90166
## Airaprae 0.012622 0.02048 0.6164 0.7143 0.0 0.92074
## Bellpere 0.012234 0.01461 0.8376 0.8571 0.0 0.93923
## Juncbufo 0.011769 0.01920 0.6131 0.5714 0.4 0.95702
## Vicilath 0.008809 0.01190 0.7406 0.2857 0.4 0.97033
## Trifprat 0.007942 0.01288 0.6164 0.2857 0.4 0.98234
## Comapalu 0.007235 0.01503 0.4814 0.0000 0.4 0.99328
## Empenigr 0.004449 0.01112 0.4002 0.2857 0.0 1.00000
## Chenalbu 0.000000 0.00000 NaN 0.0000 0.0 1.00000
## Cirsarve 0.000000 0.00000 NaN 0.0000 0.0 1.00000
## Permutation: free
## Number of permutations: 0Reduccion de la dimensionalidad Nonmetric Multidimensional Scaling (NMDS)
MDS<- metaMDS(dune)## Run 0 stress 0.1192678
## Run 1 stress 0.188969
## Run 2 stress 0.1183186
## ... New best solution
## ... Procrustes: rmse 0.02026853 max resid 0.06495039
## Run 3 stress 0.1183186
## ... New best solution
## ... Procrustes: rmse 2.958027e-05 max resid 9.263908e-05
## ... Similar to previous best
## Run 4 stress 0.1183186
## ... Procrustes: rmse 2.613872e-05 max resid 8.369925e-05
## ... Similar to previous best
## Run 5 stress 0.119268
## Run 6 stress 0.1183186
## ... Procrustes: rmse 6.250777e-05 max resid 0.0002073138
## ... Similar to previous best
## Run 7 stress 0.1183186
## ... Procrustes: rmse 5.640112e-05 max resid 0.0001621743
## ... Similar to previous best
## Run 8 stress 0.1183186
## ... Procrustes: rmse 4.484322e-05 max resid 0.0001322764
## ... Similar to previous best
## Run 9 stress 0.119268
## Run 10 stress 0.1192679
## Run 11 stress 0.1183186
## ... New best solution
## ... Procrustes: rmse 1.192512e-05 max resid 3.876647e-05
## ... Similar to previous best
## Run 12 stress 0.1192679
## Run 13 stress 0.1183186
## ... Procrustes: rmse 4.020905e-05 max resid 0.0001239373
## ... Similar to previous best
## Run 14 stress 0.1183187
## ... Procrustes: rmse 4.619113e-05 max resid 0.0001331066
## ... Similar to previous best
## Run 15 stress 0.1183186
## ... Procrustes: rmse 2.818575e-05 max resid 9.388134e-05
## ... Similar to previous best
## Run 16 stress 0.1183186
## ... Procrustes: rmse 3.707039e-05 max resid 0.0001104109
## ... Similar to previous best
## Run 17 stress 0.1192683
## Run 18 stress 0.1192678
## Run 19 stress 0.1183187
## ... Procrustes: rmse 5.320446e-05 max resid 0.0001689346
## ... Similar to previous best
## Run 20 stress 0.1192679
## *** Solution reachedplot(MDS)
Este grafico es muy feo, vamos a hacer un ggplot!
DF<- MDS$points %>% as.data.frame() %>% bind_cols(dune.env)
ggplot(DF, aes(MDS1, MDS2, col=Management))+
geom_point()+
labs(title = "Diferencias entre tipos de manejo")+
theme_bw()
Vemos que los puntos de BF y HF se agrupan, lo que indica que hay diferencias en estos grupos frente a todo el conjunto.
En cuanto al tipo de uso tenemos
ggplot(DF, aes(MDS1, MDS2, col=Use))+
geom_point()+
labs(title = "Diferencias entre tipos de uso")+
theme_bw()
En este caso vemos que los puntos de los grupos estan dispersos, luego no se evidencia diferencia.
Transformaciones en MDS
Cuando tenemos comunidades balanceadas en numero de especies la reduccion de dimensionalidad no tiene mucho problema, pero que pasa si tenemos una o varias especies dominantes con una gran cantidad de registros? en este caso hay que hacer transformaciones para balancear los pesos antes de realizar el analisis de reduccion de dimensionalidad.
dune2<- dune
dune2[1,1]<-100 #cambiamos el valor de las observaciones del sitio1 para la primera especie por 100
MDS2<- metaMDS(dune2)## Square root transformation
## Wisconsin double standardization
## Run 0 stress 0.1052259
## Run 1 stress 0.1042872
## ... New best solution
## ... Procrustes: rmse 0.02063417 max resid 0.0620581
## Run 2 stress 0.1052259
## Run 3 stress 0.1052278
## Run 4 stress 0.1052278
## Run 5 stress 0.2279892
## Run 6 stress 0.1052259
## Run 7 stress 0.1052278
## Run 8 stress 0.1042407
## ... New best solution
## ... Procrustes: rmse 0.005194691 max resid 0.01624794
## Run 9 stress 0.1042872
## ... Procrustes: rmse 0.00519584 max resid 0.0163146
## Run 10 stress 0.1052259
## Run 11 stress 0.1052259
## Run 12 stress 0.1691151
## Run 13 stress 0.1042874
## ... Procrustes: rmse 0.005198679 max resid 0.01626654
## Run 14 stress 0.1052278
## Run 15 stress 0.1781424
## Run 16 stress 0.1052259
## Run 17 stress 0.1052259
## Run 18 stress 0.1042872
## ... Procrustes: rmse 0.0051978 max resid 0.01633038
## Run 19 stress 0.1042407
## ... Procrustes: rmse 2.112988e-05 max resid 7.045355e-05
## ... Similar to previous best
## Run 20 stress 0.1042872
## ... Procrustes: rmse 0.005196663 max resid 0.01632158
## *** Solution reachedLa funcion metaMDS realiza una transformacion de raiz cuadrada y una estandarizacion doble de Wisconsin de forma automatica. Se recomienda hacer las transformaciones de forma manual para tener claridad metodologica en el proceso.
Que variables influencian la diferencia entre comunidades?
Vamos a ajustar un modelo que nos permita identificar las variables que causan diferencias entre las comunidades.
Fit<- envfit(MDS, dune.env, perm=999)
plot(MDS)
plot(Fit)
plot(Fit, p.max=0.05, col="red")
En rojo podemos ver las variables que son significativas