Soil life in reconstructed ecosystems: Initial soil food web responses after rebuilding a forest soil profile for a climate change experiment

Abstract

Disrupting ecosystem components, while transferring and reconstructing them for experiments can produce myriad responses. Establishing the extent of these biological responses as the system approaches a new equilibrium allows us more reliably to emulate comparable native systems. That is, the sensitivity of analyzing ecosystem processes in a reconstructed system is improved by excluding the period when observed phenomena are primarily responses caused by establishing the experiment rather than effects of imposed treatments; achieved by determining the extent of any pulse of activity caused by preparatory procedures. A native forest soil was physically disrupted when it was collected, sieved, and then rebuilt in lysimeters in a controlled-environment study evaluating the influence of elevated atmospheric CO 2 concentration and elevated atmospheric temperature on the reconstructed soil that was planted with Douglas-fir ( Pseudotsuga menziesii Mirb. Franco) seedlings. Generally, soil food web populations responded in two phases during the exposure as indicated by preliminary evaluation of the 4.5-year dataset. Also, previous work indicated that relatively elevated soil CO 2 effluxes occurred during the first phase, suggesting that food web populations may have responded to carbon sources made available when the soil was harvested and its profile reconstructed in the lysimeters. Results are presented for bacterial and fungal biomass, numbers of protozoa and nematodes to gain insight on whether the first phase responses are attributable to the acute stress of physically disrupting the soil. We found clear relationships between changes in predator and prey populations. A prominent spike for many of the food web populations occurred the year after the climate exposures began. Except for total bacterial biomass and total fungal-hyphae biomass, overall food web responses generally were unrelated to treatments. It appears that initial food web population responses were related to increased availability of soil carbon caused by establishing the experiment. Our results provide insights into determining the length of time to maintain reconstructed forest ecosystems before responses are observed related to experimental treatments. It appears that as long as 3 years elapsed before the soil food web appeared to recover from the acute physical disturbance; 1 year of recovery prior to commencing the climate exposures to allow the soil to rest after it was reconstructed, plus approximately the first 2 years of maintaining the climate treatments. Accounting for consequences of such periods of adjustment is critical for forecasting whether comparable natural ecosystems will be net sources or sinks of elevated concentrations of atmospheric CO 2.