I am not quite clear on this question. But you might use allometric equations from the literature. But the best was is to harvest above and below ground, seperate into different components (leaves, roots, etc.) and then calulate relations hips with each comonent something easy to measure such a baseline diameter or height and then apply those regressions to the shrubs you want to sample.
We determined the shrub biomass of Larrea divaricata (Zygophyllaceae) in central Argentina applying destructive methods (stems, leaves, roots) and then we used models to estimate biomass using variables of easy measurement as: height, crown diameter, etc. with the lowest estimated error.