AQUARISK 3.2
Upgrade Notes
Upgrade notes from Version 2.5 to Version 3.2
The output tables that are displayed after input of water quality or dose response data now include the statistics derived for both distribution forms. Right-clicking in the header of either table gives a drop down menu enabling the user to select an option of viewing either a full or a summary version of table, and enables the user to select which columns they would like to print in their output report or to sort the table by any column values.
It should be noted that the method of calculating a user defined hazardous concentration (HC) value has been changed (for lognormal distributions of dose response data). This follows advice, from consultants on statistics, that the linear regression method does not give accurate results of the uncertainty of distribution parameters. Therefore, a bootstrapping method is used to estimate a user defined HC value (HCx,y – see below). Bootstrapping is also used for estimating HC5,50, HC5,95 and HCx,y values when the Burr III distribution is selected for the dose response data. The calculation of the p statistic is likely to be changed shortly.
When you use the Tools menu to view or print Water Quality or Dose Response Data Graphs you will notice that the figures now include both distribution forms. The blue is the log-normal distribution and the red curve is Burr III. The information below the figure has been amended to include statistics and parameters for both distributions.
The user may now also save the figures in wmf format for insertion into reports. This is done using the Tools menu available on the CDF plot pages.
When the user selects Derive criteria from dose response data a new dialog box appears that allows the user to a) decide whether to use the existing TOX files and b) to set site-specific species protection levels.
With respect to the TOX files, the user should select No (at least) for the first ERA on any site. This will allow the user to establish the data sets and filtering required for the site in question. This action is also necessary on the first occasion for those users who may be upgrading their code from an earlier version. If the user selects Yes, an error message will be generated if TOX files have not been created previously. Alternatively the user runs the risk of assessing the water quality data against inappropriate dose-response information.
With respect to the option of setting site-specific criteria for ecological risk in terms of the level of species protection (higher is more protective) and uncertainty (higher is more precautionary). These values should ideally be agreed by stakeholder input.
Assuming the user has selected No in response to the option to use the existing TOX files, a browse box opens to select the appropriate data file as in the previous version. Once selected, a new dialog box appears with three options. It is highly recommended that, at this stage, any data filtering to improve the site- or species- specificity of the dose-response data be carried out on all toxicant worksheets as described in the previous handbook. After filtering, the first option is to create selected TOX files. If this option is taken the process is the same as in the previous version of AQUARISK. The second option, missing, is to create TOX files only for contaminants that do not have TOX files (i.e. where previous TOX files do not exist in the AQUARISK directory). This is recommended for subsequent ERAs on any site where assessments have previously been applied (except when upgrading from a previous version of the code, see below) and where new toxicants have been included. Please note that the dose-response data for the new toxicants should be filtered using the same criteria that were applied for the earlier assessments. The third option is to create TOX files for all contaminants. This option is recommended for first time users and initial uses of upgrade codes.
Due to the multiple sampling required to derive Burr III parameters, establishing the TOX files may take some time on the first occasion the code is used. The larger the data set the longer the time required. The time will also depend upon the number of toxicants and the processing speed of the computer. Subsequent uses should proceed much more quickly. The brief delay at this time is to allow calculation of the site-specific species protection criterion for each toxicant.
There is now an option in the drop down Tools menu on the main page for Selection of Distributions Fitted. This option allows the user to quickly change the output of any of the subsequent actions to reflect the use of the different distribution forms. However, both the Uncertainty Analysis option on the drop down Tools menu and the Estimate reduction required in chemical concentration required to achieve a specified level of average % species affected on the main page are unavailable when Burr III distribution form has been selected for either the water quality data or the dose response data. This is because of the excessively long calculation time required for estimating these values by convolution of Burr III distributions. In addition, the Estimate average % of species affected output remains the same as in previous versions when using log-normal distributions. However, when using Burr III two different methods are used and reported accordingly. These methods are numerical integration and Monte Carlo sampling.
An additional feature in the Estimate average % of species affected output is given at the base of the table. This now provides an estimate of the total percentage of species affected due to the impact of all toxicants assuming proportional additivity.
The Users Guide and Technical Manual, both in pdf format, have been updated to reflect these changes. The in-built help files have also been updated.