We use the transport theory to simulate three-dimensional radiation distribution in a vegetation canopy of a small area (ca. 0.1-0.3 ha). This theory is based on two contradictory assumptions (Ross, 1981). On the one hand, the model resolution have to be so high that input variables for the transport equation can approximate the given forest stand with necessary degree of accuracy. On the other hand, the transport theory is based on the assumption that Beer's law can be locally applied to plant canopies that is valid for sufficiently large volumes filled with phytoelements. This sets a limit to the resolution and to the predicting accuracy, not only of our model, but also of any other using Beer's law. The aim of our paper is to estimate these limits as a function of input variables. A detailed analysis of input variables (canopy structure, optical properties of foliage elements and soil, radiation input at the canopy boundary) and of their effect on the radiative field underlie our investigations. A comparison of our three-dimensional simulation results with field measurements is also included in our paper, not only to test the model, but also to illustrate the specification of a model resolution and the accuracy of predicted radiative field in a real small heterogeneous experimental site. The forest albedo is an important ecological variable characterising the forest scattering capacity. To measure this variable, two hemispherical sensors are usually mounted above the forest canopy. The first one records a downward energy flux from the atmosphere, and the second, an upward irradiance reflected by the forest. The ratio of their responses is usually interpreted as the forest albedo. As an example, the model is used to quantify an inadequacy of this interpretation by simulating both the sensor response and the three-dimensional distribution of the radiation reflected, and by comparing these results with measurements in the field. (C) 1997 Elsevier Science B.V.
