Phosphorus (P) availability in terrestrial ecosystems depends on soil age, climate, parent material, topographic position, and biota, but the relative

We collected soils from both Icacos and Bisley, in different topographic positions, and analyzed them for phosphorus content.

Abstract from the paper listed below Mage & Porder 2013: https://www.doi.org/10.1007/s10021-012-9612-5

Phosphorus (P) availability in terrestrial ecosystems depends on soil age, climate, parent material, topographic position, and biota, but the relative
importance of these drivers has not been assessed. To ask which factor has the strongest influence on long- and short-timescale metrics of P availability, we sampled soils across a full-factorial combination of two parent materials [quartz diorite (QD) and volcaniclastic (VC)], three topographic positions (ridge, slope, and valley), and across 550 m in elevation in 17 sub-watersheds of the Luquillo Mountains, Puerto Rico. VC rocks had double the P content of QD (600 vs. 300 ppm; P < 0.0001), and soil P was similarly approximately 29 higher in VC-derived soils (P < 0.0001). Parent material also explained the most variance in our two other longtimescale metrics of P status: the fraction of recalcitrant P (56% variance explained) and the loss of P relative to parent material (35% variance explained), both of which were higher on VC-derived soils (P < 0.0001 for both). Topographic position explained an additional 10–15% of the variance in these metrics. In contrast, there was no parent material effect on the more labile NaHCO3- and NaOH-extractable P soil pools, which were approximately 2.59 greater in valleys than on ridges (P < 0.0001). Taken together, these data suggest that the relative importance of different state factors varies depending on the ecosystem property of interest and that parent material and topography can play sub-equal roles in driving differences in ecosystem P status across landscapes.

Lab Analysis:

Soils were collected in Luquillo National Forest, Puerto Rico, in July 2010.
We collected soils from 16 subwatersheds in a full factorial combination of 2 forest types (colorado, tabonuco) and 2 parent materials (quartz diorite, volcaniclastic)
In each subwatershed we dug 9 soil pits to a depth of 80cm (0-20, 20-50, 50-80). Three pits were located in ridge, three in slope, and 3 in valley positions.
This yielded a total of 432 soil samples, which were air dried at the University of Pennsylvania. A subsample was then shipped to Brown University
At Brown, equal mass of the three replicates from a given topographic position within a subwatershed were combined to make a composite sample (e.g. all three 0-20cm depth ridge samples from site COOX-1 were combined).
We analyzed for P fractions using a modified Hedley fractionation, extracting sequentially with 0.5N NaHCO3 and 0.1N NaOH. Each extract was measured for inorganic P on a westco smartchem 200 analyzer.
Each extractant was also digested with persulfate and reanalzyed for total P, organic P was determined by difference between P in the persulfate digest and P in the undigested extractant.
Every 10th sample was run in triplicate to ensure reproducibility.
Each of the composite samples was also digested by lithium borate flux fusion and run by XRF to determine total elements, and by ICP-MS to determine trace elements. This work was done
by ALS Chemex (SParks, NV).
The data presented on the next page are concentrations of different P forms per gram oven dry weight of soil.

Tau P refers to the loss (<0) or gain (>0) of P relative to parent material, using the P concentrations of soil and parent material referenced to niobium (see Porder and Chadwick, 2009). Tau = -0.5 represents a 50% loss
of P relative to the parent rock.