Integrating estimates of tree root mass predicted with ground penetrating radar and allometry

Ground penetrating radar (GPR) operated in reflection mode may be used to estimate lateral root biomass in forests. The technique has been very useful for quantifying belowground biomass and accounting for carbon in silivicultural studies. In general, surface-based GPR cannot detect fine roots (<2 mm diameter), vertical taproots, decayed roots or separate roots by species. This presents challenges when integrating GPR-based assessments of lateral roots (between trees) and below-stump biomass estimates (directly below trees) modeled from stand inventory data (e.g. diameter, height) as there may be overlap between the approaches. To support ongoing research in longleaf pine (Pinus palustris mill.) ecosystems, we scanned 11 longleaf pine trees aged 15 to 79 years with GPR and compared the results to excavations. A 16 m2 area around each tree was surveyed with 1500 MHz antenna via 9 parallel lines, 0.5 m apart with the tree located in the center. A root biomass map was made for each tree. The size of the excavated pit around each tree was calculated from the linear relationship between tree basal area and pit size e.g. younger trees < 25 yo 1.0 to 1.5 m2, older, larger trees 1.5 to 4.0 m2. Excavated roots were classified as lateral (roughly perpendicular to stem) and taproot (vertical roots) then weighed after drying. The area of the excavated pit was noted on the root biomass map and the mass detected in that area was calculated and compared to the mass of lateral roots. The proportion of roots “missed” by GPR was negligible for small trees ~10 cm diameter at breast height (DBH), increased to 87% for the largest tree (54 cm DBH). Fortunately, the proportion of longleaf pine lateral roots detected by GPR can be predicted using an exponential decay function fitted with tree DBH. The assumption that GPR detects all lateral roots may be valid for small trees (<10 cm DBH), though underestimation of mass is expected with larger trees.