Blecker, S. W., R. L. McCulley, O. A. Chadwick, and E. F. Kelly. 2006. Biologic cycling of silica across a grassland bioclimosequence. Global Biogeochemical Cycles 20:1.
Problem statement: cycling of Si through biomass in grassland systems remains unknown. This cycling must be measured and be considered in any estimation of Si weathering. Previous studies have examined Si cycling in various forest ecosystems, but never in grasslands. This is somewhat surprising, considering that grasslands are considered to be the highest producers of phytoliths in the world. Also, how does biologic cycling affect mineral Si weathering and export to watersheds?
Goals: To determine if Si stored in biomass varies as a function of climate. To measure how quickly Si is cycled through biomass. To see how much of an impact grasslands have on Si cycling and storage.
Study area: an east-west “bioclimosequence” from W Missouri to NE Colorado. The dominant vegetation transitions from tall-grass prairie in the east to short-grass in the west.
Methods: Soil samples to the base of the C horizon at eight study sites. Soils were described. Water samples were taken to determine dissolved Si. Soil phytoliths were extracted. Plant samples were also taken, and phytoliths were extracted from them.
Results: Biogenic Si was usually highest in the topsoil, and decreased downward. This is similar to organic carbon content. Plants absorb dissolved Si (monosilicic acid) through their roots in the lower topsoil and B horizon. Upon death, the Si is deposited in the upper topsoil as phytoliths. Through time the phytoliths begin to dissolve in the topsoil and leach into the B horizon, where they are re-absorbed by roots. The authors claim that more phytoliths are entrained in short-grass sites, due to lower precipitation, and therefore lower dissolution. This is somewhat counterintuitive, since tall-grass prairie sites have higher overall phytolith production. However, these results agree with other studies, namely Alexandre et al. (1997), which found high dissolution rates in tropical rainforest systems. It seems plausible then, that lower moisture ecosystems could have a higher phytolith residence time in the soil Si pool.
Limitations: While the authors did use a bioclimosequence, in which sites became progressively drier as one heads west, they did not account for many factors which could have dramatically influenced their results. Some of these factors include:
Differing soil types. Some soils may have high Si amounts, while others may have lower. This will certainly affect the biogenic Si turnover rate. For example, a soil which has a large pool of readily dissolvable mineral Si would presumably have higher soil phytolith residence times. On the other hand, an Si depleted soil would be expected to have higher amounts of biogenic Si dissolution. Therefore, differing soil types and Si pools must be accounted for, or at the very least, every attempt must be made to minimize these differences.
Non-compatible temporal resolutions. Some of the sample sites may be transitory; that is, the type of vegetation found at a given site may vary from year to year. For example, any ecologist knows that marginal species habitat zones are the most susceptible to climate change. Ecosystems which may be mixed-grass at the present may have been tall-grass only a few years ago. Since the temporal resolution of phytolith assemblages can be up to 200 years, the phytolith record would not reflect this change. Thus conflicting results would emerge: mixed-grass vegetation coupled with a low soil phytolith pool (which is most likely present due to the past tall-grass vegetation). Therefore, attempts must be made to determine the medium- to long-term vegetation. This can be accomplished from a number of techniques, the least of which would be to analyze the very phytoliths used in this study (although some may see that as circular reasoning).
An arbitrary classification technique (short-, mixed-, and tall-grass prairie). Look at figure 3c, which shows phytolith abundances throughout the soil profiles of the mixed-grass prairie sites. The three sites vary substantially, even though they are supposedly in the same ecosystem. The authors conclude that the mixed-grass sites have the highest soil phytolith entrainment rate based on this graph. While they certainly could be right, I believe more research is necessary before a sound conclusion can be made.
A small data set. Eight study sites were used in this study, which is probably enough. However, a greater number of samples could have been taken from some of the soil profiles. For example, the Wah-Kon-Tah site had only three soil samples taken. I don’t think one should measure the overall phytolith concentration throughout the entire soil profile based on only three data points.
Improper biogenic Si extraction technique. My own research has shown that a large percentage of central Great Plains soil is made up of Oligocene aged volcanic glass. For phytolith analysis, this can be a problem. The standard procedure for the removal of phytoliths from a soil is to place the entire sample in a “heavy” liquid, with a density of 2.3 g cm-1. Most minerals have a density around 2.65 g cm-1, while phytoliths are usually less than 2.3. When the sample and the heavy liquid are centrifuged, the phytoliths will float to the top of the liquid, since they are lighter. The minerals will sink to the bottom. In this way, the phytoliths are effectively separated from the minerals. Unfortunately, volcanic glass tends to be the same density as phytoliths, making the two very difficult to separate. For any study which is attempting to measure the total amount of biogenic Si (phytoliths plus any other Si which has been used by organisms), traditional floating techniques really aren’t appropriate. First, floatation techniques are very time consuming and expensive. Second, it is almost impossible to completely separate minerals from biogenic Si, even if no volcanic glass is present. Third, traditional methods cannot separate small biogenic Si particles (< 5 µm) from clay and fine silt particles. For these reasons, any quantitative study should look to other methods, such as biogenic Si dissolution using NaOH or Na2CO3 (see Saccone 2005).
Conclusions: I think this study is a good first step for understanding the Soil Si cycle in grasslands. However, I believe that a larger data set should be used. Also, more robust attempts should be made to minimize those variables which could influence the results. In other words, sample sites should be as similar as possible, with respect to soil type, elevation, etc.