The Palmer-Barber method is an elastic analysis that assumes each subgrade layer behaves as a linear, isotropic, homogeneous material. Real subgrades are none of these things: they are non-linear (stiffness decreases at higher stress levels), often anisotropic (horizontal and vertical stiffnesses differ), and highly variable across the plan area of a building or pavement structure. Engineers should keep the following limitations in mind when using this calculator.

• Input uncertainty. The primary source of uncertainty is the modulus input for each layer. An error of a factor of 2 in the natural subgrade M_R — a realistic range for CBR-based conversion — propagates directly into the k-value results. For CBR values above 10%, the linear Heukelom-Klomp formula overestimates stiffness and the power-law AASHTOWare Pavement ME form should be used instead. For subgrades with CBR above 20%, direct laboratory resilient modulus testing per ASTM D3999 is recommended rather than relying on any CBR conversion.
• Spatial variability. The method uses a single k-value for the entire slab or pavement structure. In reality, the subgrade is rarely uniform across the plan area of a large warehouse or along a roadway. Areas near drainage ditches, utility trenches, column footings, or transitions between cut and fill zones can have dramatically different stiffnesses. Wherever significant variability is suspected, multiple investigation points should be used and the design k should be based on the characteristic lower-bound value, not the average.
• Long-term consolidation. The elastic k-value captures the initial elastic response of the subgrade but does not account for long-term consolidation. Under sustained loading from heavy rack systems, cohesive subgrades — clays and silts — will consolidate over months or years, producing settlement that is not reflected in the k-value. For slabs supporting static rack loads above approximately 11,000 lb (5,000 kg) per leg, a consolidation settlement check is recommended in addition to the structural slab design.
• Layer assumptions. The method assumes full friction (fully bonded interfaces) between all layers and perfectly horizontal, infinite layers. Real granular subbases may exhibit partial slip at the subgrade interface, and layer thickness variations, slope, or lateral boundaries are not accounted for. Similarly, the load radius is fixed at 15 in. (38.1 cm), the standard plate diameter, which is a reasonable approximation for pavement design but may not match the actual contact geometry of specialized loading equipment.
• The 2× dynamic-to-static factor. The 2× factor between dynamic and static k — used to convert the Palmer-Barber output to a static k suitable for design methods calibrated to plate load tests — is an empirical observation, not a derived quantity. Field measurements consistently support this ratio across a wide range of subgrade types, but it can vary from roughly 1.5× to 3× depending on soil type, moisture condition, and loading rate. Engineers working with design methods that specify a static k should verify which testing basis was used to calibrate the method.
• Cement-treated subbases. For cement-treated subbases with a very high modulus relative to the underlying subgrade, the Palmer-Barber method tends to underpredict the k increment at thin subbase thicknesses by 10–15% compared to empirical plate load test data. Performance improves at greater thickness or with higher subgrade stiffness. For projects where this range of conditions is expected, the empirical PCA or FAA composite k-value tables provide a useful cross-check.

Despite these limitations, the Palmer-Barber method achieves average errors comparable to the disagreement between independent empirical datasets (PCA vs. FAA), and it handles multi-layer conditions that the older correlation tables cannot address at all. It is well suited to routine design practice when used with appropriate engineering judgment and project-specific soil investigation.