Preprints
https://doi.org/10.5194/bg-2021-78
https://doi.org/10.5194/bg-2021-78

  01 Apr 2021

01 Apr 2021

Review status: this preprint is currently under review for the journal BG.

A Comparative Assessment of Gap-filling Techniques for Ocean Carbon Time Series

Jesse M. Vance1, Kim Currie2, John Zeldis3, Peter Dillingham4, and Cliff S. Law1,5 Jesse M. Vance et al.
  • 1Department of Marine Science, University of Otago, Dunedin, 9016, New Zealand
  • 2National Institute of Water and Atmospheric Research – University of Otago Centre for Oceanography, Dunedin, 9016, New Zealand
  • 3National Institute of Water and Atmospheric Research, Christchurch, 8011, New Zealand
  • 4Department of Statistics, University of Otago, Dunedin, 9016, New Zealand
  • 5National Institute of Water and Atmospheric Research, Wellington, 6021, New Zealand

Abstract. Regularized time series of ocean carbon data are necessary for assessing seasonal dynamics, annual budgets, interannual variability and long-term trends. There are, however, no standardized methods for imputing gaps in ocean carbon time series, and only limited evaluation of the numerous methods available for constructing uninterrupted time series. A comparative assessment of eight imputation models was performed using data from seven long-term monitoring sites. Multivariate linear regression (MLR), mean imputation, linear interpolation, spline interpolation, Stineman interpolation, Kalman filtering, weighted moving average and multiple imputation by chained equation (MICE) models were compared using cross-validation to determine error and bias. A bootstrapping approach was employed to determine model sensitivity to varied degrees of data gaps and secondary time series with artificial gaps were used to evaluate impacts on seasonality and annual summations and to estimate uncertainty. All models were fit to DIC time series, with MLR and MICE models also applied to field measurements of temperature, salinity and remotely sensed chlorophyll, with model coefficients fit for monthly mean conditions. MLR estimated DIC with a mean error of 8.8 umol kg−1 among 5 oceanic sites and 20.0 ummol kg−1 among 2 coastal sites. The empirical methods of MLR, MICE and mean imputation retained observed seasonal cycles over greater amounts and durations of gaps resulting in lower error in annual budgets, outperforming the other statistical methods. MLR had lower bias and sampling sensitivity than MICE and mean imputation and provided the most robust option for imputing time series with gaps of various duration.

Jesse M. Vance et al.

Status: final response (author comments only)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on bg-2021-78', Adrienne Sutton, 23 Apr 2021
    • AC1: 'Reply on RC1', Jesse Vance, 02 Jul 2021
  • RC2: 'Comment on bg-2021-78', Anonymous Referee #2, 19 May 2021
    • AC2: 'Reply on RC2', Jesse Vance, 02 Jul 2021

Jesse M. Vance et al.

Jesse M. Vance et al.

Viewed

Total article views: 460 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
343 105 12 460 2 4
  • HTML: 343
  • PDF: 105
  • XML: 12
  • Total: 460
  • BibTeX: 2
  • EndNote: 4
Views and downloads (calculated since 01 Apr 2021)
Cumulative views and downloads (calculated since 01 Apr 2021)

Viewed (geographical distribution)

Total article views: 342 (including HTML, PDF, and XML) Thereof 342 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 26 Jul 2021
Download
Short summary
Long-term monitoring is needed to detect changes in our environment. Time series of ocean carbon have aided our understanding of seasonal cycles and provided evidence for ocean acidification. Data gaps are inevitable, yet no standard method for filling gaps exists. We compared eight methods to understand the impact of difference approaches when assessing seasonality and annual carbon budgets. We suggest the regression approach used here be considered when determining the best practices.
Altmetrics