Impacts of fire and fire surrogate treatments on forest soil properties: a meta-analytical approach.

The soils underlying the 12 Fire and Fire Surrogates Network include six soil orders and >50 named soil series. Across the network, pretreatment soils varied from 3.7 to 7.1 in pH, and exhibited ranges of twofold in bulk density, fourfold in soil organic C (SOC) content, 10-fold in total inorganic N (TIN), and 200-1000-fold in extractable Ca and K. Nonmetric multidimensional (NMS) ordination of pretreatment soil conditions arrayed the FFS sites along gradients of pH/base cation status, net N transformation rates, bulk density, and SOC. At the network scale, mineral soil exposure was significantly greater in fire-only (mean of 9.2%) and mechanical + fire (5.0%) treatments than in the controls (1.5%) during the first posttreatment year, and this persisted through the later sampling year (second through fourth year, depending on site) in the fire-only treatment (fire 4.1%, control 1.1%). Bulk density was not affected significantly at the network scale. TIN concentrations during the first posttreatment year increased after all three manipulative treatments, but this effect did not persist to the later sampling year. Neither SOC content nor soil C:N ratio was affected by any of the treatments at the network scale. At the individual site scale, the combined mechanical + fire treatment produced more significant site x treatment x year effects than did the fire-only or mechanical-only treatments, though in most cases even the statistically significant differences produced by the manipulative treatments were modest in magnitude. Ordination of first-year standardized effect sizes produced no discernable separation of the three manipulative treatments but did separate the three sites with the greatest fire severity (based on proportional fuel consumption) from the majority of the network sites, with changes in pH, TIN, SOC content, and soil C:N ratio correlating most strongly with this separation. Ordination of the effect sizes from the later sampling year produced somewhat clearer separation of treatments than did the first-year ordination, though fewer sites were represented in this second ordination. Overall, the network-wide effects of the FFS treatments on soil properties appear to have been modest and transient.