The measurement of forests is essential for monitoring and predicting the role and response of the land surface to global climate change. Globally consistent and frequent measurements can only be made by satellites; unfortunately many current system’s measurements saturate at moderate canopy densities and are not directly related to forest properties, requiring tenuous empirical relationships that are insensitive to many of the Earth’s most important, Carbon rich forests. Lidar (laser radar) is a relatively new technology that offers the potential to make direct measurements of forest height, vertical density and, when ground based, explicit measurements of structure. In addition measurements do not saturate until much higher forest densities. In recent years there has been much interest in the measurement of forests by lidar, with a number of airborne and terrestrial and one spaceborne lidar developed. Measuring a forest leaf by leaf is impractical and very tedious, so more rapid ground based methods are needed to collect data to validate satellite derived estimates. These rapid methods are themselves not directly related to forest properties causing uncertainty in any validation of remotely sensed estimates. This thesis uses Monte Carlo ray tracing to simulate the measurement of forests by full waveform lidars over explicit geometric forest models for both above and below canopy instruments. Existing methods for deriving forest properties from measurements are tested against the known truth of these simulated forests, a process impossible in reality. Causes of disagreements are explored and new methods developed to attempt to overcome any shortcomings. These new methods include dual wavelength lidar for correcting satellite based measurements for topography and a voxel based method for more directly relating terrestrial lidar signals to forest properties.