gcb2511-sup-0001-FigureS1.tifimage/tif1484KFigure S1. Estimates of Net Ecosystem Exchange (NEE) for; (panel a) the forward model run with 500 random samples from the prior parameter and stock estimates, (panel b) the EnKF and (panel c) the Metropolis algorithm. The analyses shown are for L = 10 years, σ = 0.58 g C m−2 day−1 and Rg = 0.6. Dots mark the noisy observations. For clarity, uncertainties on the observation are not shown. Solid lines represent the mean NEE estimates. Grey shading indicate the NEE estimate range for the 5th and 95th percentiles.
gcb2511-sup-0002-FigureS2.tifimage/tif1282KFigure S2. The effect of observation noise on parameter estimation. Panels a[RIGHTWARDS ARROW]q show the 17 DALEC-D parameters, on the top half is the EnKF, on the bottom half is the Metropolis algorithm. Panels r[RIGHTWARDS ARROW]w show the six initial carbon pools required for the Metropolis algorithm. The dot represents the mean value, with the error bars indicating the 5th and 95th percentiles. The true parameter values are indicated with red dotted horizontal lines. In the case of the Metropolis algorithm, the prior mean is indicated with a blue line.
gcb2511-sup-0003-FigureS3.tifimage/tif1019KFigure S3. As for Figure S2, but showing the effect of observation noise out on final stock estimation.
gcb2511-sup-0004-FigureS4.tifimage/tif1289KFigure S4. As for Figure S2, but showing the effect of data gaps on parameter estimation.
gcb2511-sup-0005-FigureS5.tifimage/tif1025KFigure S5. As for Figure S2, but showing the effect of data gaps on final stock estimation.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.