Implement timezone fix for summarizing uv to dv

[lindsayplatt]

Similar to the groundwater visualization, summarizing UV values to DV likely has issues with timezones being different from the site's local timezone. See the fix implemented here and the reason why we probably haven't noticed it in this repo here.

Unfortunately, this likely extends to the historic data that we use in this repo, which is summarized from UV to DV in the national-flow-observations due to how we are summarizing UV to DV. I imagine there is minimal difference, since there would only be a shift of 6 values between days, especially for the historic values.