Planning a Seasonal Mosquito Surveillance Program

Begin with three to five questions in plain language. Where is activity concentrated relative to prior seasons? Did the habitat work at these four locations change anything observable? Are we servicing the sites we committed to servicing? Can we substantiate coverage to the board?

Each question implies requirements: a comparable baseline, a defined before-and-after window, a completeness measure, a documented site inventory. Written down in advance, these become the design constraints. Discovered afterward, they become the reasons the data cannot be used.

A surprising share of analysis failures trace to site identity. Locations get renamed, duplicated, merged, or moved a short distance without a record. When identity is unstable, history cannot be assembled, and a site’s own baseline — the most valuable comparison available — becomes unusable.

Establish a canonical list with stable identifiers, recorded coordinates, and a documented procedure for what happens when a location moves or is retired. Assign someone to own it. This is unglamorous and pays for itself in the first review cycle.

Decide what interval each site needs to be serviced at for its records to be comparable across the season, then check that against staffing. Where the two conflict, reduce the number of sites rather than degrading the interval everywhere.

A program with thirty consistently serviced sites produces usable evidence. A program with seventy inconsistently serviced sites produces a larger archive that supports fewer conclusions. This trade is uncomfortable to make in spring and obvious in hindsight.

Every field should have a stated definition, an allowed set of values or units, and a rule for what to record when the observation cannot be made. Ambiguity here produces the divergence between crews that destroys comparability, and it does so invisibly.

Include condition fields — weather during the interval, equipment state, access issues, vegetation changes — as first-class entries rather than optional notes. These are what make a low count interpretable later.

Schedule at least one review during the season, early enough that correction is still possible. The purpose is not analysis; it is data quality. Are fields being completed? Are crews interpreting definitions the same way? Are servicing intervals holding? Are any sites systematically missed?

Errors caught in July can be corrected. The same errors discovered in November are permanent features of the dataset. This checkpoint is the highest-leverage hour in the season plan.

Name the specific deliverables the season will produce: a coverage summary, a site-level comparison against baseline, a completeness measure, a list of prioritized locations with reasons. Knowing the outputs determines whether the inputs are sufficient.

It also removes the end-of-season scramble where staff reconstruct summaries under deadline from records that were never organized for that purpose.

Site inventory, form definitions, servicing schedule, mid-season review, and final reporting each need a named owner. Shared ownership of documentation reliably produces no documentation.

Ownership should survive staff turnover, which means writing the procedures down rather than relying on the person who has always done it. Programs lose more institutional knowledge to undocumented routine than to any analytical shortcoming.