User Tools

Site Tools


en:rarefaction_examples

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
en:rarefaction_examples [2019/03/22 22:04]
David Zelený
en:rarefaction_examples [2019/03/26 01:37]
David Zelený [Example 1: Comparing forest diversity along elevation standardized to sample area, number of individuals and sample completeness]
Line 59: Line 59:
   * for ''​datatype = "​incidence"'':​ ''​site''​ and ''​S.obs''​ the same as above; ''​T''​ is the number of plots within each locality (25 subplots in our case), U is the overall number of species incidences within locality (i.e. the sum of incidences of individual species, where incidence = presence of the species in one subplot). ''​Q1'',​ ''​Q2'',​ ''​Q3'',​ ... ''​Q10''​ are numbers of species occurring in only 1, 2, 3, ... 10 subplots within each locality (unique species, duplicate species, etc.).   * for ''​datatype = "​incidence"'':​ ''​site''​ and ''​S.obs''​ the same as above; ''​T''​ is the number of plots within each locality (25 subplots in our case), U is the overall number of species incidences within locality (i.e. the sum of incidences of individual species, where incidence = presence of the species in one subplot). ''​Q1'',​ ''​Q2'',​ ''​Q3'',​ ... ''​Q10''​ are numbers of species occurring in only 1, 2, 3, ... 10 subplots within each locality (unique species, duplicate species, etc.).
  
-(note that both abundance and incidence data are based on the same original dataset, which contains number of individuals surveyed within each of 25 10x10-m subplots; for abundance data, individuals of each species have been summed across all 25 subplots within the locality, while for incidence data, only presences-absences of species within the subplots (i.e. not the number of their individuals) were considered and summed ​accross ​all 25 subplots within the locality).+(note that both abundance and incidence data are based on the same original dataset, which contains ​number of individuals surveyed within each of 25 10x10-m subplots; for abundance data, individuals of each species have been summed across all 25 subplots within the locality, while for incidence data, only presences-absences of species within the subplots (i.e. not the number of their individuals) were considered and summed ​across ​all 25 subplots within the locality).
  
-Let's focus on abundance-based data first (''​hp.abund''​). You can see that localities quite remarkably ​differ ​in numbers of individuals (''​n''​),​ with the lowest number in FT (Feng-Tien, 232 individuals,​ low elevation) and highest in WJ (Wu-Jie, 1731 individuals,​ middle elevation). The numbers of species somehow copy the number of individuals (the correlation between ''​n''​ and ''​S.obs''​ in ''​DataInfo (hp.abund)''​ is 0.7: ''​cor (DataInfo (hp.abund)$S.obs,​ DataInfo (hp.abund)$n)''​). This may be suspicious; what if the middle elevation localities are diverse simply because they have a higher density of individuals per fixed sampled area?+Let's focus on abundance-based data first (''​hp.abund''​). You can see that localities ​are quite remarkably ​different ​in numbers of individuals (''​n''​),​ with the lowest number in FT (Feng-Tien, 232 individuals,​ low elevation) and highest in WJ (Wu-Jie, 1731 individuals,​ middle elevation). The numbers of species somehow copy the number of individuals (the correlation between ''​n''​ and ''​S.obs''​ in ''​DataInfo (hp.abund)''​ is 0.7: ''​cor (DataInfo (hp.abund)$S.obs,​ DataInfo (hp.abund)$n)''​). This may be suspicious; what if the middle elevation localities are diverse simply because they have a higher density of individuals per fixed sampled area?
  
 To make sure that this is not the case, let's standardize the data to fixed number of individuals. We can first draw the rarefaction curves to see differences between individual localities: To make sure that this is not the case, let's standardize the data to fixed number of individuals. We can first draw the rarefaction curves to see differences between individual localities:
Line 93: Line 93:
 6   PL 232 interpolated 0.970 25.229 15.268 11.799 6   PL 232 interpolated 0.970 25.229 15.268 11.799
 7   GY 232 interpolated 0.979 17.965 ​ 7.373  5.236 7   GY 232 interpolated 0.979 17.965 ​ 7.373  5.236
-</code>+</file>
 Note that, for simplicity, in the further comparisons I decided to ignore confidence intervals (''​conf = NULL''​ in ''​estimateD''​). If you do the comparison seriously, you should, however, consider them; in that case, the output of ''​estimateD''​ will be slighlty more complex. Note that, for simplicity, in the further comparisons I decided to ignore confidence intervals (''​conf = NULL''​ in ''​estimateD''​). If you do the comparison seriously, you should, however, consider them; in that case, the output of ''​estimateD''​ will be slighlty more complex.
  
Line 137: Line 137:
 D_area <- DataInfo (hp.abund, datatype = '​abundance'​)$S.obs D_area <- DataInfo (hp.abund, datatype = '​abundance'​)$S.obs
 D_est <- cbind (D_area, D_individuals,​ D_coverage) D_est <- cbind (D_area, D_individuals,​ D_coverage)
-rownames (D_est) ​<- D_area ​<- DataInfo (hp.abund, datatype = '​abundance'​)$site+rownames (D_est) <- DataInfo (hp.abund, datatype = '​abundance'​)$site
 D_est D_est
 </​code>​ </​code>​
Line 157: Line 157:
 </​code>​ </​code>​
 {{:​obrazky:​hp_abund_barplot_comparison.png?​direct|}} {{:​obrazky:​hp_abund_barplot_comparison.png?​direct|}}
 +
 +
 Note that it make no sense to compare diversities of individual plots across different standardisations (e.g. richness of FT standardised to area, individuals and coverage), but it make sense to compare them within the same standardisation (e.g. FT standardised to number of individuals with YYH standardised to number of individuals). From the barplot it is clear that the rank of localities according to their richness changes after standardisation;​ e.g., after standardisation to sample coverage, the YYH (Yuan-Yang-Hu,​ the plot close to famous Yuan-Yang lake, 鴛鴦湖, perhaps the foggiest locality in Taiwan) became species poorer than FT (Feng-Tien, lowland subtropical forest), although in the original data YYH is richer than FT, perhaps due to remarkably higher number of individuals surveyed in YYH (1551 ind.) than in FT (232 ind.). Note that it make no sense to compare diversities of individual plots across different standardisations (e.g. richness of FT standardised to area, individuals and coverage), but it make sense to compare them within the same standardisation (e.g. FT standardised to number of individuals with YYH standardised to number of individuals). From the barplot it is clear that the rank of localities according to their richness changes after standardisation;​ e.g., after standardisation to sample coverage, the YYH (Yuan-Yang-Hu,​ the plot close to famous Yuan-Yang lake, 鴛鴦湖, perhaps the foggiest locality in Taiwan) became species poorer than FT (Feng-Tien, lowland subtropical forest), although in the original data YYH is richer than FT, perhaps due to remarkably higher number of individuals surveyed in YYH (1551 ind.) than in FT (232 ind.).
  
en/rarefaction_examples.txt · Last modified: 2019/03/26 01:37 by David Zelený