Tropical soils are high in irons and minerals that have the ability to bind P, preventing leaching, and release it under certain conditions . Recent research has questioned how much of this sequestered P may be available to plants or microbes, or even regulated by large scale ecosystem changes in redox capacity liberating P from mineral associations . Potassium is also derived from weathering of bedrock parent material with mineral K being slowly released by weathering, while exchangeable K is highly mobile and easily swapped between plants and clays . Limitation of K in lowland tropical soils is especially likely as Ultisols and Oxisols tend to be high in clays which release exchangeable K quickly but do not fix K within mineral associations . Tropical soils have been functionally categorized as variable-charge soils in which the mobility of cations like K+ and NH4+ may exceed the mobility of anions such as nitrate when soil pH is low . Highly weathered soils may be deficient in K as well as P, with K limitation potentially varying systematically with soil acidity. High precipitation and weathering rates of most tropical lowland forests, along with high K mobility, suggest K is likely to be a limiting nutrient. Ecosystem limitation by K is an important consideration not only due to its potential to be easily leached but also its effects on growth and productivity. A study of root growth and litter decomposition in three Amazonian forests gives early evidence of K limitation in tropical ecosystems, with K leaching consistently more rapidly than other macro-nutrients under all experimental conditions . Measurements of nutrient cycling by stem flow, through fall and litter fall in a temperate swamp forest showed K was rapidly cycled with up to 93% of seasonal soil K inputs occurring via through fall leaching . Rapid cycling of K has also been observed in tropical agroforest species with litter leaching of K significantly exceeding that of P and N in all three species examined .

The importance of K as a limiting nutrient is also indicated by experimental evidence of plant responses to K availability: A meta-analysis of 38 K fertilization experiments of 26 tree species from temperate and tropical forests showed that 69% of studies reviewed observed a significant positive responses in growth to fertilization . In addition,plant benches review of five known water-shed scale studies measuring K+ runoff showed significant decreases in K+ runoff during the growing season, with patterns of K+ runoff tending to mirror those of NH4+ . Such studies suggest that K may be commonly deficient in many ecosystems, with K addition frequently producing positive growth results, while large scale fluxes of K+ exhibit similar patterns to NH4+ indicate that similar mechanisms of biotic control may govern ecosystem-scale cycling of both nutrients. The broad importance of K in ecosystem function is becoming increasingly evident, however, its role in ecosystem nutrient cycling dynamics remains poorly explored in most ecosystems, and specifically in the tropics. Most of our understanding of nutrient demand and limitation in tropical ecosystems has been indirectly inferred from measurements of above ground plant responses to fertilization or community scale measurements of canopy nutrient content. Stand-level trends in leaf nutrient abundance have been associated with changes in nutrient resorption proficiency, and total soil nutrient availability, suggesting above ground nutrient concentrations are reflective of overall ecosystem nutrient limitation . Remote sensing of canopy nutrient concentrations and measurements of ecosystem C flux have largely been regarded as a barometer for nutrient limitation on productivity, with above ground nutrient and C dynamics presumed to reflect those below ground. Fertilization studies of tropical trees have shown that concentrations of N and P in leaves and twigs closely resemble the concentrations observed in fine roots , indicating that above ground measures of tissue nutrient concentrations are an expression of below ground nutrient availability; however, this approach is unable to distinguish between demand for the uptake of nutrient and the utilization of previously acquired nutrients. Measurements of plant root nutrient uptake offer a potentially direct insight into plant nutrient demand or acquisition capacity.

Studies examining nutrient uptake have largely been conducted in hydroponically grown agricultural species with little to no in situ mature trees studied in the tropics. While results from these studies have been central to our understanding of nutrient roles in physiology they do not accurately represent nutrient cycling dynamics in established ecosystems. Nutrient demand should be reflected through changes in the rate of active nutrient uptake by roots when soil fertility is altered. Nutrients that are limiting may be individually targeted for uptake, in cases where only one nutrient limits total plant function, or non-specific nutrient uptake should increase in order to access more total nutrients when one or more nutrients may functionally increase productivity. Changes in nutrient uptake would expectedly occur in conjunction with changes in growth rates, tissue nutrient concentrations, or alterations of physiological productivity and biomass allocation. Direct measurements of root nutrient uptake may allow for a more complete insight into nutrient demand and limitation than only above ground measurements. A solution-depletion method can be used to directly measured root nutrient uptake, using roots connected to the tree immersed in nutrient solution to measure net changes in nutrient concentrations over time . For example, this method has shown seasonal variation in N uptake in mature mid-latitude red spruce and sugar maple, with increased N uptake observed during the growing season and a pattern of greater nitrate uptake at the start of the growing season and greater ammonium uptake at the end . Root uptake of N has also been observed to vary diurnally in mature Spruce and Beech trees with both species exhibiting greater uptake rates during daytime as well as a preference for NH4+ uptake with greater soil availability . Species specific preferences for NH4+ over NO3- have also been observed across coniferous and deciduous species including oaks, beeches, hemlocks, and maple . Differential uptake of NH4+ and NO3- has been reported across multiple mid-latitude species with in-situ uptake rates ranging from 10-15 umol/g dry weight*hr NH4+ and 1-5 umol/g dry weight*hr NO3- .

Only one study of mature loblolly pine has so far observed K uptake, which exhibited similar seasonal trends as NH4+ with greater uptake in the early growing season . Measurements of nutrient influx over short timescales afford the possibility of measuring subtle responses of nutrient demand to environmental conditions. Variations of this method are also readily applied to laboratory grown plants allowing for a uniform method to compare the nutrient uptake characteristics of seedlings to those of mature trees. A CO2 fertilization experiment of red maple and sugar maple seedlings showed the addition of CO2 did not affect N uptake rates of sugar maple,rolling bench while red maple seedlings exhibited a significant increase in NH4+ uptake but not NO3- . A similar experiment with ponderosa and loblolly pine seedlings found nitrate uptake increased in both species with CO2 fertilization while ammonium uptake decreased significantly in ponderosa seedlings . Soil temperature and fertility were also observed to influence N uptake of spruce and beech seedlings with NO3- uptake significantly inhibited by low temperatures and high NH4+ availability, while NH4+ uptake rates increased linearly with soil temperature across a range of 20° C . Studies using this method have shown that root nutrient uptake is highly dynamic and relatively species specific, however is also responsive to a range of environmental variables that allow for greater quantification of the role of nutrient limitation in ecosystem function. Nearly all studies of mature tree root-nutrient uptake have been conducted in mid-latitude forests, largely with an emphasis on N uptake, leaving a gap in our understanding of P and K dynamics, and tropical ecosystems. Kuntze was selected as the focal tree species for this study as it was common across all of the plots. Tetragastris panamensis is a shade tolerant, mid and upper canopy species with heights ranging from 10-35 m. In this species, root biomass accounts for roughly 25% of total plant C storage in sampled plants from a Costa Rican lowland rainforest . Here, one mature T. panamensis was selected per plot, and roots were traced from the stem, gently excavated to depths of ~20cm, kept in-tact and attached to the main stem, and then exposed to nutrient solutions for nutrient dilution experiments . In plots where roots were inaccessible due to downed logs, steep slopes, or confounding trees nearby, a substitute tree was used in a previously sampled plot, thus selecting a second individual at least 5 meters away from the individual tree to be used as an independent replicate. One plot from each of the C, N, P, K, NP, and NPK treatments required resampling of a second tree within an already sampled plot; no plot was sampled for more than two trees and each treatment had at least three plots sampled. For the purposes of this study, individual trees are used as replicates . We used a well-established nutrient dilution method to assess nutrient uptake rates, with modifications as noted . Nutrient solutions consisting of distilled water, ammonium , nitrate , potassium , and phosphorus were mixed together at three different concentrations in order to assess nutrient uptake with variation in nutrient concentration. Concentrations included: 25 umol/L NH4+, 15 umol/L NO3-, 10 umol/L P, 20 umol/L K; 125 umol/L NH4+, 75 umol/L NO3-, 50 umol/L P, 100 umol/L K; 250 umol/L NH4+, 150 umol/L NO3-, 100 umol/L P, 200 umol/L K. Three roots from each individual tree were exposed to each of the three concentration levels, and a water-only concentration to collect root exudates and assess leakage of nutrients from roots. Ammonium was added to distilled water as ammonium chloride , nitrate as sodium nitrate, and P and K as dipotassium phosphate . Mixing of solutions was done in batches every 3-6 days to avoid nitrification over extended periods. Roots were placed into 50 mL centrifuge tubes filled to 40ml for each concentration. Three replicate vials for each of the three concentration levels were used for each tree, plus water replicates, for a total of 12 vials per tree. Twelve roots from each tree were gently excavated, rinsed with DI water, and inserted into sample vials.Roots were exposed to solution for 3–4 hours after which roots were clipped at the point of entry to the solution, removed, and dried at 60 ℃ until weight stabilized and weighed. The solutions were filtered through pre-rinsed Whatman Grade 1 filters , then frozen for analysis on the same day they were collected. One control vial at each concentration was also included without inserting a root to assess any background effects or variation in nutrient concentrations. Vials were weighed before and after solutions were added to measure changes in solution mass and try to quantify passive nutrient acquisition through water uptake. Water-only solutions were used to assess water intake and measure background root leakage of nutrients and carbon. Changes in solution mass within vials are not significantly affected by, or correlated with, any treatment or covariate. Vials were weighed on a scale accurate to three digits; mean solution weights and standard deviations before and after the experiment for each treatment were measured .Gross uptake rates for individual roots were calculated as the net uptake of each root minus the average rate of nutrient leakage for the sampled tree. Net uptake rates are calculated as the difference in nutrient concentration between the nutrient-control vial, and the experimental vial, divided by the dry root weight in grams and experiment duration . Rates of nutrient leakage and carbon exudation were calculated similarly for water-only vials. The rates of nutrient leakage from the three water-only exposed roots were averaged to produce an average rate of leakage for each nutrient for the tree. Gross uptake rates were used in our analysis to correct for the effects of root leakage. The final units for gross uptake are in micro-moles per liter per gram dry weight per hour . Results from the simplified model are presented alongside full model results for both datasets.