Abstract
A compartmental carbon model is presented in which there are parameters quantifying the transfer between various carbon pools. These parameters are estimated using two CO2 data sets (elevated and ambient) from the Duke Forest FACE data set. A quadratic objective criterion is used to compare closeness of fit of system output CO2 flux to the observed CO2 flux data. Adjoint equations are developed based on the system of difference equations approximating the solution of the model equations, the data, and the criterion. Parameter estimation algorithms are developed using minimization techniques. It is found that the estimated parameters obtained using the two different data sets agree closely.
Original language | English (US) |
---|---|
Pages (from-to) | 101-120 |
Number of pages | 20 |
Journal | Applied Mathematics and Computation |
Volume | 130 |
Issue number | 1 |
DOIs | |
State | Published - Jul 25 2002 |
Externally published | Yes |
Keywords
- Adjoint method
- Carbon dioxide cycle
- Deconvolution
- Estimation
ASJC Scopus subject areas
- Computational Mathematics
- Applied Mathematics