Data File(s) | Type | Description | Action |
---|---|---|---|
model_comparisons.zip (4.88 KB) | ZIP Archive (ZIP) | Statistical intermediate that compares the number of times a given model was within 2AIC points of the lowest AIC score for that scale. The model that is selected as having the lowest AIC score most often is ultimately selected as the 'best' model and used to study variable effects over scale. Naming scheme is: scale of analysis, species code, "model_comparison". Parameters in the .csv files of the folder: Column Name,Column Description [Include meaning of any codes or flags used in data column as well as detection limits.],Units of measurement,missing data/no data value best_model ,GAM formula for the model being compared,no units,no missing values count ,the number of instances over the course of the all permutations of buffer and edge at this analysis scale that this model was within 2 points of having the lowest AIC score ,count ,no missing values r.sq ,the average adjusted GAM r.sq value over all scales. ,no units,no missing values mean_buffer,the mean buffer value for all instances where this model was selected. ,meters ,no missing values | |
model_lists.zip (17.97 KB) | ZIP Archive (ZIP) | List of models that were determined to be within 2 AIC points of the best model for each permutation of edge and buffer. Parameters of the .csv files: Column Name,Column Description [Include meaning of any codes or flags used in data column as well as detection limits.],Units of measurement,missing data/no data value edge,"distance from the marsh edge in meters used to define edge habitat (1,3,or 5)",meters ,no missing values buffer,radius of habitat region around the sample site used to aggregate habitat data. ,meters ,no missing values sp,species code for the current analysis,no units,no missing values best_model,model formula ,no units,no missing values aic,aic score for the model ,no units,no missing values r.sq ,the average adjusted GAM r.sq value over all scales. ,no units,no missing values | |
output_over_scale.zip (14.37 KB) | ZIP Archive (ZIP) | Tables that show selected model (as seen in the top_model.txt file corresponding to the species and analysis) outputs over all scales in the analysis. naming convention follows: analysis scale, "_", species code, "output_over_scales.csv". Parameters of the .csv files: Column Name,Column Description [Include meaning of any codes or flags used in data column as well as detection limits.],Units of measurement,missing data/no data value edge,"distance from the marsh edge in meters used to define edge habitat (1,3,or 5)",meters ,no missing values buffer,radius of habitat region around the sample site used to aggregate habitat data. ,meters ,no missing values sp,species code for the current analysis,no units,no missing values best_model,model formula ,no units,no missing values r_sq ,the average adjusted GAM r.sq value over all scales. ,no units,no missing values dev_ex,deviance in response variable explained by the model,no units,no missing values moran_P,P value for Moran's I test ,no units,no missing values cov,column indicating presence of covariance between variables. -Inf indicates no covariance,no units,no missing values p.edge_perc,p value for % edge habitat ,no units,no missing values p.edge_l.mangrove,p value for % mangrove habitat,no units,no missing values p.land_water_ratio,p value for %land ,no units,no missing values p.man_lwr,p value for interaction of % mangrove and % land ,no units,no missing values p.man_edge,p value for interaction of % mangrove and % edge,no units,no missing values | |
spatial_analysis_across_scales.R (135.27 KB) | R Script | main code file, takes inputs from linked datasets and produces outputs shown here | |
species_aic.zip (119.04 KB) | ZIP Archive (ZIP) | AIC scores for all models tested at every permutation for every species. naming convention is: species code, "_", scale of analysis, "edge", edge distance (m), "buf", buffer distance (m). | |
topmodel.zip (2.48 KB) | ZIP Archive (ZIP) | Text files with the formula of the top model chosen for the analysis for each species at each analysis scale. Naming convention is: analysis scale, species code, "_topmodel". |
Files
Type: ZIP Archive (ZIP)
Description: Statistical intermediate that compares the number of times a given model was within 2AIC points of the lowest AIC score for that scale. The model that is selected as having the lowest AIC score most often is ultimately selected as the 'best' model and used to study variable effects over scale. Naming scheme is: scale of analysis, species code, "model_comparison". Parameters in the .csv files of the folder: Column Name,Column Description [Include meaning of any codes or flags used in data column as well as detection limits.],Units of measurement,missing data/no data value best_model ,GAM formula for the model being compared,no units,no missing values count ,the number of instances over the course of the all permutations of buffer and edge at this analysis scale that this model was within 2 points of having the lowest AIC score ,count ,no missing values r.sq ,the average adjusted GAM r.sq value over all scales. ,no units,no missing values mean_buffer,the mean buffer value for all instances where this model was selected. ,meters ,no missing values
Type: ZIP Archive (ZIP)
Description: List of models that were determined to be within 2 AIC points of the best model for each permutation of edge and buffer. Parameters of the .csv files: Column Name,Column Description [Include meaning of any codes or flags used in data column as well as detection limits.],Units of measurement,missing data/no data value edge,"distance from the marsh edge in meters used to define edge habitat (1,3,or 5)",meters ,no missing values buffer,radius of habitat region around the sample site used to aggregate habitat data. ,meters ,no missing values sp,species code for the current analysis,no units,no missing values best_model,model formula ,no units,no missing values aic,aic score for the model ,no units,no missing values r.sq ,the average adjusted GAM r.sq value over all scales. ,no units,no missing values
Type: ZIP Archive (ZIP)
Description: Tables that show selected model (as seen in the top_model.txt file corresponding to the species and analysis) outputs over all scales in the analysis. naming convention follows: analysis scale, "_", species code, "output_over_scales.csv". Parameters of the .csv files: Column Name,Column Description [Include meaning of any codes or flags used in data column as well as detection limits.],Units of measurement,missing data/no data value edge,"distance from the marsh edge in meters used to define edge habitat (1,3,or 5)",meters ,no missing values buffer,radius of habitat region around the sample site used to aggregate habitat data. ,meters ,no missing values sp,species code for the current analysis,no units,no missing values best_model,model formula ,no units,no missing values r_sq ,the average adjusted GAM r.sq value over all scales. ,no units,no missing values dev_ex,deviance in response variable explained by the model,no units,no missing values moran_P,P value for Moran's I test ,no units,no missing values cov,column indicating presence of covariance between variables. -Inf indicates no covariance,no units,no missing values p.edge_perc,p value for % edge habitat ,no units,no missing values p.edge_l.mangrove,p value for % mangrove habitat,no units,no missing values p.land_water_ratio,p value for %land ,no units,no missing values p.man_lwr,p value for interaction of % mangrove and % land ,no units,no missing values p.man_edge,p value for interaction of % mangrove and % edge,no units,no missing values
Type: R Script
Description: main code file, takes inputs from linked datasets and produces outputs shown here
Type: ZIP Archive (ZIP)
Description: AIC scores for all models tested at every permutation for every species. naming convention is: species code, "_", scale of analysis, "edge", edge distance (m), "buf", buffer distance (m).
Type: ZIP Archive (ZIP)
Description: Text files with the formula of the top model chosen for the analysis for each species at each analysis scale. Naming convention is: analysis scale, species code, "_topmodel".