Validating records that are not supported by a voucher

I am currently working my way through iRecord data for 2017. Some of it is relatively straightforward, especially where I know the recorder and his/her abilities and techniques. But, what do you do with records from somebody you don't know who has submitted a record that might be plausible but you know could easily be wrong?

Several records gave me this conundrum today. I had the usual problem with Syrphus ribesii being used for all yellow and black striped Syrphus. For these, unless there is a photo that I can work from, I mark all with a query because I've seen so many unidentifiable photographs posted as this species. I do accept records by people I know and whose techniques I am familiar with.

The bigger problems come in the case of relatively easily identified species at the wrong time of year. Two examples emerged today: three records of Volucella bomylans in August and one of Merodon equestris in late July. The problem is that they were the majority of records seen for these species in today's session and they don't tally with the normal phenology plots for either species. Fortunately, we have a way of telling whether they are well within or on the limits of believability.

This is where extracting records direct from Facebook helps. I have access to a detailed dataset for both species in 2017 all of which I have checked from photographs. In the case of V. bombylans there are 202 records and for M. equestris there are 353 records. In neither case is there much evidence of large numbers at the times of year of the questionable records, as can be seen from the accompanying graphs.

So, do I mark as plausible, or should I reject the records? What would you do?

Figure 1. Phenology of Volucella bombylans in 2017.

Figure 2. Phenology of Merodon equestris in 2017