We are currently using D-Opto dissolved oxygen sensors in marsh habitats to measure continuous seasonal changes in dissolved oxygen. Power issues and occasional sensor issues leads to gaps in our data set, which we would like to fill in order to conduct the appropriate time series analysis. Unfortunately, literature searches have yielded little information regarding methods for gap filling dissolved oxygen data. For this reason I am reaching out to the Research Gate community for advice. Thanks :)
Dear Justin,
First of all: Depending on what your ultimate objectives are, it may or may not be necessary to replace the missing data in your time series. Remember that any value you 'fill in' will necessarily be an estimate based on a number of assumptions... There are specific statistical methods to deal with incomplete or irregularly spaced time series.
Second, if you do want or need to analyze complete, regularly spaced time series, then there are many mathematical methods to do so, which can meet a wide range of requirement in terms of sophistication and performance. The issue there is what tools and techniques do you have access to or are willing to learn about?
Third, there is no reason to believe that the choice of such a 'gap filling' method should be dependent on the physical or chemical nature of the variable being measured: mathematical theory and computers are independent of thematic applications.
This being said, there is a fairly substantial literature on this topic. You may find this paper useful as an introduction to some of the methods, as well as to the literature:
http://www.atmos-chem-phys.net/11/7905/2011/acp-11-7905-2011.pdf
Cheers, Michel.
What is the frequency of measurement during normal operation? How long are the gaps? You need to provide a little detail.
- Badges
- Science topic
- Similar topics
- Phytochemistry
- Extracts