AGÕæÈË°Ù¼ÒÀÖ¹Ù·½ÍøÕ¾

Recent updates in the algorithms that generate some of our aquatic instrument data will improve published data products by fixing erroneous data. The updates will address the following issues with each of these data products at all sites where they are collected:

1.  (all Aquatic sites): A new algorithm was deployed on October 24th, 2019 that improved data quality and publication format, see NEON.DOC.002181vC (the algorithm theoretical basis document) for details on the revised processing algorithm. Data collected and processed after October 24th, 2019 should be reflected on the data portal with the December 2019 publication of late October and November 2019 data. Data collected prior to October 24th, 2019 will be re-processed and re-published over the next 6-12 months and eventually all nitrate data collected by NEON will be available in the new format. Until the re-publication of older data is complete, users can disregard the superfluous blank rows with flags every 30 seconds to one-minute and use the nitrate data published every 15 minutes for the older format.
2.  (all Aquatic sites with buoys): Software algorithm fixes have been tested and should start to be reflected in new data within the next month. A data portal news update will accompany this algorithm update. Data collected and processed after the deployment will have the correct time averaging windows. Corrections on the data portal should be fully reflected in 2020 February’s publication of 2020 January’s data. All data collected prior to the deployment will be re-processed with the correct averaging windows and re-published over the next 6-12 months. The time shift begins at approximately zero and progresses to a positive shift of approx. 1 hour at the end of 4 hours. It then resets to zero. On the daily and higher timescale, the differences between actual and reported statistics are small.
3.  (all Aquatic sites with buoys): Software algorithm fixes have been tested and should start to be reflected in new data within the next month. A data portal news update will accompany this algorithm update. Data collected and processed after the deployment will have the correct time averaging windows. Corrections on the data portal should be fully reflected in 2020 February’s publication of 2020 January’s data. All data collected prior to the deployment will be re-processed with the correct averaging windows and re-published over the next 6-12 months. We recommend using the instantaneous 1-minute temperature at specific depth in surface water data until we announce the fixes for averaged data.
4.  (all Aquatic sites): Uncertainty estimates corresponding to each data stream from the multisonde have been corrected by November 9^th, 2019 and any data collected and processed on or after that will have valid uncertainty estimates published. Data collected and processed between July 16th, 2019 and November 8th, 2019 will be re-processed and re-published over the next 6-12 months. Any data collected and processed prior to July 16^th, 2019 was not part of this issue and was published with appropriate uncertainty estimates. Data streams other than uncertainty were not impacted by this issue.
5.  (all 30-minute data, beginning Jan 1, 2018): A software algorithm fix has been identified and NEON plans to internally test and deploy the updated algorithm within the next month. A data portal news update will accompany this algorithm update. Data collected and processed after the deployment will have the correct data included in the 30-minute averages. Corrections on the data portal should be fully reflected in 2020 February’s publication of 2020 January’s data. All data collected prior to the deployment will be re-processed with the correct averaging windows and re-published over the next 6-12 months. We recommend using the 5-minute groundwater elevation data product until we announce the algorithm update.

At the time of this posting, the algorithm updates are not reflected in published data. Data are being re-processed and will be corrected. We will announce when these corrected data have been republished.