Iowa Mosquito Surveillance

Looking behind the numbers

Iowa-Mosquito.net contains various charts, but not all numbers are created equal. Depending on what dimensions you're looking at, data are represented in either a qualitative or quantitative way.

Raw data

Raw data means that we are displaying the absolute number of mosquitoes trapped at a given location at a given time. This makes sense when comparing different locations within a county. It is also logical to use this presentation method when showing a historical week-by-week histogram for a single year. It does not make sense to do this type of comparison between counties, which is why data are normalized as indicated below.

Normalized data: Weekly Average

The most important resolution in the database is the county level. In order to compare different counties, the data are normalized by dividing total numbers (raw data) by the number of individual trap locations within a county. The reason for this is that if County A has 5 trap-sites and County B has only 1, it would be easy to create the impression that A has more mosquitoes than B because there may have been more mosquitoes caught in 5 traps as opposed to 1. Therefore, to allow for cross-county comparisons, the raw data are always divided by the number of traps running.

Normalized data: Trap Index

The Trap Index (TI) is the standard representation for mosquito surveillance efforts. The TI accounts for the number of mosquitoes trapped, the duration of trapping and number of traps running (formula: # mosquitoes/(# traps x # nights running)). This is used to visualize the data when displayed on a yearly basis, because in a given year, the number of sites and duration of trapping can differ.

Proportional data

While constructing this website, it became clear that, per species, representing raw mosquito number data from year-year artificially inflated or deflated the abundance of a given species. In other words: if in year X there were twice as many individuals of species R compared to year Y, it becomes irrelevant if there were also twice as many individuals of species S. So that one can look critically at the abundance of a given species, year-year data now are represented in two ways: one way is the above-mentioned "raw data" format, and the alternative is a proportional representation that shows how the abundance of one species compares to the abundance of all species. On species pages, this view is referred to as Composition.

Log-scaled data

Because of the exponential relative increase in abundance of certain species in comparison with others, it was deemed more appropriate for visualization purposes that the data be log-scaled. An example of this data transformation type is seen in the species overview page.