Monte Carlo estimation is explored as an alternative to traditional survey sampling techniques to estimate both daily and seasonal whole-tree photosynthesis of first-year Populus clones. Several methods, known in the literature as variance-reduction techniques, are applied to the problem of estimation and compared on the basis of relative root mean squared error. Also of interest is gain in precision over the simple expansion estimator (Monte Carlo estimation in its simplest form). Variance reduction is achieved by approximating the photosynthesis curve by some known, easily integrated function. The estimators retain their unbiasedness regardless of the appropriateness of this function. The authors show how these variance reduction techniques can be used to achieve greater precision when estimating both daily and seasonal whole-tree photosynthesis. These methods may be useful alternatives to current purposive sampling methods that have the potential for bias and high error.
Monte Carlo integration with a growing number of control variates