Forest soil classification for intensive pine plantation management: “Site Productivity Optimization for Trees” system

Forest productivity and response to silvicultural treatments are dependent on inherent site resource availability and limitations. Trees have deeper rooting profiles than agronomic crops, so evaluating the impacts of soils, geology, and physiographic province on forest productivity can help guide silvicultural management decisions in southern pine plantations. Here, we describe the Forest Productivity Cooperative’s “Site Productivity Optimization for Trees” (SPOT) system which includes: texture, depth to increase in clay content, drainage class, soil modifiers (i.e., surface attributes, mineralogy, and additional limitations such as root restrictions), geologic formations, and physiographic province. We quantified the total area for each SPOT code in the native range of loblolly pine (Pinus taeda L.), the region’s most commercially important species, and used a remotely-sensed layer to quantify SPOT code areas in managed southern pine (approximately 14 million ha). The most common SPOT code in the native range is also the most planted, a B2WekoGgPD (fine loamy, shallow depth to increase in clay, well-drained, eroded, kaolinitic, granitic, Piedmont soil), spanning 1.1 million ha total, but only 12% in managed southern pine. However, the SPOT code with the greatest percentage of managed southern pine (61%; a D4PoioAmAF, spodic, deep to increase in clay, siliceous, middle Atlantic Coastal Plain, Flatwoods soil) was the 20th most common in the native range with 474,662 ha. We used machine learning and data from decades of “Regionwide” trials to assess the variable importance of SPOT constituents, climate, planting year, and N rate on site index (base age 25 years) and found that planting year was the most important variable, showing an increase of 17 cm site index per year since 1970, followed by maximum vapor pressure deficit, and precipitation. Geology was the top-ranking SPOT variable to explain site index followed by physiographic province. The Regionwide trials represent 72 unique SPOT codes (out of over 10,000 possible in the pine plantations) and approximately one million ha (or about 7% of all soils identified as supporting managed pine). To extrapolate site index values outside of the unique soil and geologic conditions empirically represented, we created a predictive model with an R2 of 0.79 and an RMSE of 1.38 m from SPOT codes alone. With this extrapolation, the Regionwide data predicts 10.5 million ha, or 74%, of all soils under loblolly pine management in its native range. Overall, this system will allow managers to assess their current site productivity, and recommend silvicultural treatments, thus, providing a framework to optimize forest productivity in pine plantations in the southeastern US.