Spearman's rank correlation analysis
In order to see whether there is a correlation between climate variables, SOI and flowering, I used a linear regression and correlation analysis. Linear regression models a relationship between two variables and can be used to predict one from the other. A correlation analysis shows the strength of the relationship between two variables that are analyzed. I performed a multiple correlation analysis analyzing flowering in each region with each of the climate variables.
I first checked my data for the assumptions of normality and equal variance. In Graph 6 you can see a histogram of the residuals of my data (the graph shows the residuals for a plot of Borneo Flowering with Precipitation but all of the variables showed similar graphs and are not shown here). Also, Shapiro test p-values showd significance for all of my residuals which means that the data anlysied is not normally distributed (not shown here). I could not transform my data as my flowering data showed a descending distribution. The most frequent observation falls on 0 GF Intensity and decreases with increasing flowering intensity. After this analysis I treated my data as non-parametric.
I first checked my data for the assumptions of normality and equal variance. In Graph 6 you can see a histogram of the residuals of my data (the graph shows the residuals for a plot of Borneo Flowering with Precipitation but all of the variables showed similar graphs and are not shown here). Also, Shapiro test p-values showd significance for all of my residuals which means that the data anlysied is not normally distributed (not shown here). I could not transform my data as my flowering data showed a descending distribution. The most frequent observation falls on 0 GF Intensity and decreases with increasing flowering intensity. After this analysis I treated my data as non-parametric.
Graph 6: Histogram of the Residuals
First I wanted to see what correlations there are between El Niño and climate variables in the three regions under study. Since my data did not meet the needed assumptions of normality, I used Spearman’s rank correlation test in R software for non-parametric data. I ran a pair-wise correlation analysis comparing all the climate variables for the three regions and SOI. I also performed a correlation analysis for monthly climate data and SOI to see whether stronger correlations were observed in a particular month.
In order to answer my research question, I ran a correlation analysis on the flowering intensity data for the three regions and climate variables to see what correlations and possible relationships can be established. Again, since my flowering data did not have a normal distribution, I used a Spearman’s rank correlation analysis to run pair-wise correlations with my data.
Next, I wanted to see whether there was a correlation between weather, SOI and flowering in a particular month. I conducted a Spearman’s correlation test for flowering and monthly climate and SOI data. Finally, I plotted linear regressions for the climate variables that showed the highest correlation in a particular month with flowering for the three regions studied.
In order to answer my research question, I ran a correlation analysis on the flowering intensity data for the three regions and climate variables to see what correlations and possible relationships can be established. Again, since my flowering data did not have a normal distribution, I used a Spearman’s rank correlation analysis to run pair-wise correlations with my data.
Next, I wanted to see whether there was a correlation between weather, SOI and flowering in a particular month. I conducted a Spearman’s correlation test for flowering and monthly climate and SOI data. Finally, I plotted linear regressions for the climate variables that showed the highest correlation in a particular month with flowering for the three regions studied.
