|
Hi, My inventory has a nested sub-plot design with sub-plots of various sizes where different sized trees are measured. I have specified the plot as the base unit of aggregation in the sampling design. Several variables (e.g. primary disturbance type and topographic information) are recorded at the plot level. I would like to be able to show plot counts by various categories in Saiku. I thought I would be able to make a plot count variable in the plot entity using When I try to make this variable I am not allowed and get the error message "You can not add a quantitative variable to the sampling unit entity." I presume that this is because plot is the base unit of aggregation in my inventory. Can you please suggest a workflow to allow me to show plot counts for categorical variables in Saiku? Many thanks, Jonathan |
|
Dear Jonathan, Calc cannot show quantitative variables for the base unit level. Please see discussion and solution with R script: plot (or base unit) results are written into a CSV file: https://www.openforis.support/questions/778/new-rscript-variable-in-calc-21-a3-on-plot-level?page=1&focusedAnswerId=803#803 Just a question: could your base unit be a subplot? In that case, in the sampling design you should select "Cluster" sampling where your cluster is 'plot', and not select "Apply cluster sampling only for variances". And then you could get the number of plots via Saiku too. regard, Lauri |
|
Dear Lauri, Thanks for this response and the clarification regarding extracting plot-level data to CSV using R. It is not totally clear to me whether the base unit should be the plot or a sub-plot. My inventory uses cluster sampling with clusters located on a systematic grid. Each cluster contains five plots and the plots are far enough apart to be considered independent sampling units. When defining the sampling design in Calc I selected plot as the base unit, selected "Cluster" sampling, and selected "Apply cluster sampling only for variances". As I mentioned previously each plot contains nested sub-plots of various dimensions each measuring different sized trees and stored within a different entity in the database schema (sub-plot A, B, C, and seedlings). I like the idea of specifying a sub-plot as the base unit so that I can display plot data via Saiku. However, the question then becomes which sub-plot should I select as the base unit and whether the estimator will provide valid results when configured in this manner. I greatly appreciate your expert advice on this matter. Many thanks, Jonathan |
|
Dear Jonathan, thank you for clarification. I think that in your case you have done all right because sub-plots are split into different entities. In that case you cannot select one particular sub-plot as a base unit. If you would have just one sub-plot table in Collect, then you could do it, and forget the clusters because plot are independent sampling units. About getting reliability estimates: if plots are considered independent sampling units, the (relatively easy) method is to write the plots level result variables (volume/biomass/carbon?) into a CSV file (as shown in the linked answer), and compute the variances etc. simply in Excel or R. regards, Lauri |
|
Dear Lauri, Thank you for this clarification. I will export the plot-level summaries and compute the variance in R. Thanks, Jonathan |