Doubtful Daymet data in Mexico

[piatscheck]

Hi,

(I believe I'm in the right forum to ask Daymet-related questions. My apologies if this is not the case)

I'm a biologist looking to obtain daily site-specific weather estimate in Baja California (Mexico) for the years 2011-2015. I used both the single pixel extraction tool (https://daymet.ornl.gov/single-pixel/) and the daymetr package to obtain data from 18 different locations across the peninsula (9 study sites and 9 nearby weather stations).

See the image attached here obtained with the single pixel extraction tool. These are daily temperature estimation at the ___location of a weather station in the town of San Bartolo. It looks to me that the temperature data cannot be trusted after the year 2012: the variation between days increases suddenly and the estimates reach unlikely values (maximum temperature below 0°C) in this area. I also noticed that the value would not pass 50°C, and this is not a display problem, data obtained with daymetr show the same pattern.

The pattern is the same across all the other locations in Baja California, Mexico, that I am interested in. I found daily maximum (not even minimum) temperature estimates at other study sites being around -20°C which is impossible to happen without killing a lot of frost-sensitive plants in the area, which I know didn't happened.

While I was looking for weather station data on the NOAA website, I found that data for many weather stations in Mexico are unavailable after the year 2012.

So my questions are:
- Do you know why the Daymet meteorological observations behave like this past 2012?
- Is this related to the unavailability of observations from weather stations in Mexico (seen with NOAA)?
- Are there other platforms to obtain Mexican weather stations past 2012?
- Is there any alternative to obtain daily site-specific near-surface maximum temperature data in Mexico?

Thanks,
Sincerely
Finn Piatscheck

[ORNLx - mthornton]

Hello Fin,
Thank you for your analysis and comments about Daymet data in Mexico.
I'm the Daymet production lead at the ORNL DAAC. The Baja area of Mexico is a known issue that has been corrected in what will be the next release of Daymet; due out later this year. You are correct that the errors are attributed to a sharp decrease in weather stations from Mexico; going from several thousand stations to only a few hundred in more recent years. We have improved the regression methods in the Daymet algorithm to appropriately handle these challenges in observation inputs.

Having corrected estimates later this year is probably not helpful to your current needs. If you'd like, please contact me directly and I can extract or provide you with the data that you need. thorntonmm@ornl.gov

Best wishes,
Michele Thornton

-------------------------------------
Michele Thornton
Oak Ridge National Laboratory
Distributed Active Archive Center
Environmental Sciences Division/Climate Change Science Institute
thorntonmm@ornl.gov
https://daac.ornl.gov

[piatscheck]

Hi Michele,

Thanks so much for the explanations and the offer. I will email you soon.

Best wishes,
Finn