Calculate LUMaR (and other index) scores from macroinvertebrate samples taken in the Melbourne region

LUMaR calculator: using bugs to assess stream health

Calculate LUMaR (and other index) scores from macroinvertebrate samples taken in the Melbourne region

Waterway Ecosystem Research Group Melbourne Waterway Research-Practice Partnership

This page allows you to upload macroinvertebrate data files (collected from the Melbourne region), calculate LUMaR scores from them, and (with planned additions to the tool) explore the taxa driving any trends of interest in the data. (SIGNAL score calculation is also planned)
If you are proficient in R, and want to explore the macroinvertebrate models that underlie LUMaR as well as independently perform the analyses that this page does, then you can install the melbstreambiota package.
To install the package type the following into your R console.
install.packages('', repos = NULL, type = 'source')
Please note that the package is 663 Mb, and can take ~15 min to install once it is downloaded on a powerful computer. (Once it is installed, it is quick to load and use).
Associated resources for using the package:
Manual, which details the usage of all package functions, and contents of all package data;
Vignette, which illustrates the functions of the package with worked examples.

...But if you just want to calculate lumar scores for your existing data and get the website to do the calculation, upload your data or use an example dataset
Upload an excel (.xls or .xlsx) with 2 worksheets: the first containing your sample details and the second containing the associated macroinvertebrate data*.
Your data or example data?

*LUMaR requires pairs of samples collected using standard Rapid Bioassessment sampling methods (see this document by EPA Victoria).
Data from the 2 samples need to be combined. (Counts can be summed or converted to presence-absence [if a taxon was present in 1 or both samples, count = 1; if it was absent from both, count = 0. The counts will be ignored as LUMaR uses only presence-absence.) Samples must be taken within 1 year of each other: if they were sampled between January and June, consider them Autumn samples; between July and Dec, Spring. For sample pairs taken on different dates, set the date as the average of the two dates.
The first table needs to be a list of the sample pairs with at least the following fields (others will be ignored):
  1. samppr - label for each pair of RBA samples - results will be returned using this identifier;
  2. nspring - number of spring samples (only valid values are: 0, 1, or 2);
  3. nriff - number of riffle samples (only valid values are: 0, 1 or 2);
  4. process - 'field-sort' or 'lab-sort';
  5. date - formatted as a date in Excel;
  6. subc - subcatchment ID of reach sampled, from Melbourne Water's DCI subcatchment layer. You can find the DCI subcatchment ID for many sites on the site map of the macroinvertebrate assemblage prediction tool. If this doesn't meet your needs, e-mail Chris Walsh for information on how to gain access to the shapefile.
The second table needs to contain the combined macroinvertebrate data for each samppr listed in the first table:
The table can be in one of two forms:
  1. A cross-tab table with sampprs in columns. The first column contains EPA/AUSRIVAS bugcodes** for each taxon, and subsequent columns have field names equivalent to the list of sampprs in the sample-pair table, with each row populated by presence-absence or count data - absences should = 0, abundance values will be converted to 1s
  2. A 2-column table with fieldnames samppr and bugcode (any count field will be ignored, so make sure there are no zero values in the list. If a bugcode is listed against a samppr, it is assumed to be present in the sample pair.)

    **Here is a link to the bugcodes for families, and this is a link to a table listing the 55 families used to calculate LUMaR. If your dataset contains taxa that have been identified to genus or species, and you have them listed by their bugcode, or if you have multiple entries for the same family in any one samppr, they will be combined to a single record per family per sample pair. (The 8-character taxon code is truncated to the first 4 characters, which identify families.)

Once uploaded, the provided raw data will be shown here.

Then check the next panel for how the data is manipulated before calculating LUMaR. If you are happy with that go to the final tab to calculate and download LUMaR results.

Input File Name

Sample-pair data

Macroinvertebrate data

Before calculating LUMaR, please check that changes to the input data are ok.

Sample-pair table with predictor variables for LUMaR calculation

Cleaned bugtable with the following changes applied (if any)

web analytics