Review of Derry et al. (2005)

Derry, L.A.; Kurtz, A.C.; Ziegler, K.; Chadwick, O.A. 2005. Biologic control of terrestrial silica cycling and export to watersheds. Nature 433.

Background and problem statement. Germanium/silicon (Ge/Si) ratios in streams are always lower than the primary bedrock from which they drain. Kurtz et al. (2002) found that the excess Ge is stored in secondary minerals. Therefore, higher Ge/Si values should be found in older soils. However, dissolved Ge and Si in rivers can come from both primary and secondary minerals, which can skew the interpretation. Primary minerals will export more Si to streams than secondary minerals. Thus, these primary minerals should have a high dissolved total Si (termed [Si]) and a low Ge/Si. Conversely, secondary minerals (clays) should have a low [Si] and a high Ge/Si. The authors term this the Murnane, Stallard, Froelich (MSF) model, after the authors who proposed the idea. This model can be seen in Fig. 1, or in my recreation above. Since the authors are concerned with two sources of Si in the study streams (primary vs. secondary mineral weathering), they have termed the high Ge/Si and low [Si] (derived from clays; old soils) sources as component 2. Low Ge/Si and high [Si] (derived from primary minerals) sources are termed component 1. Thus, armed with the two measurements and the hyperbolic curve in the figure below, it is possible to ascertain what the percentage of each component is present in the stream.

Despite this novel approach, the MSF model does not consider the role of biologic processes in the Si cycle. It assumes direct control by mineral weathering. [Si] should be controlled by the amount of weathering and Si export to streams. Likewise, Ge/Si should be controlled by the state of the weathering: older soils with more secondary minerals such as clays will be Ge enriched (which means a higher Ge/Si). But the ultimate pathway which Si takes before it reaches the stream may be highly influenced by vegetation. Plant uptake of DSi, precipitation of Si as phytoliths, and subsequent deposition of the phytoliths back into the soil may ultimately control the rate of Si dissolution and export. Further, a sizable amount of Si may be stored in the soil as phytoliths.

Goals. “To test the predictions of the MSF model and to investigate the impact of biogenic silica cycling on stream export.”

Study area. A chronosequence of Hawaiian streams.

Methods. The authors use two measurements: Ge/Si and total DSi [Si]. The authors recognized that DSi in the streams must be of two components (see above). The reason for this is simple: the Ge/Si value from fresh basalts is ~0.2 x 10-6 mol/mol. But the Ge/Si values recorded for the Hawaiian streams did not agree with the basalt Ge/Si. In fact, a mixing was recorded: Ge/Si of 0.2 x 10-6 mol/mol and [Si] >600 µM for component 1; and a Ge/Si of 2.6 x 10-6 mol/mol and [Si] ≤25 µM for component 2. As can be seen on the graph above, these components don’t quite fit together. The MSF model predicts that component 1 must be originating from bedrock weathering at the soil-regolith interface or from young soils which aren’t yet depleted of Si. On the other hand, component 2 should be originating from weathered soils, where the dissolution of clays and secondary minerals is dominant.

To test the MSF model, the authors recorded Ge/Si and [Si] in a chronosequence of Hawaiian streams. In other words, stream 1 drained a young watershed, stream 2 drained an older one, and so on. The chronosequence ranged from 0.3 to 4100 kyr (see Kurtz et al. 2002). The soils in the young watersheds have low Ge/Si and high [Si], as expected. Soil water solutions were extracted, and tested for Ge and Si.

Results. At sites older than 20 kyr, something strange was noted. Below 15 cm, DSi concentrations range from 6 to 45 µM, which is expected. The streams had concentrations of 100 µM or more, which is in contradiction of the MSF model. However, the top 15 cm of the soil did have roughly equal DSi concentrations as the streams. Thus, most of the DSi in the old soils is found in the topsoil, even though DSi is easily leached. Other studies had found this oddity as well, but had attributed the high Si values in the topsoil to dust import.

At the young sites (0.3 kyr), DSi is high: 200-600 µM. This is expected from young soils. However, the DSi in the topsoil is extremely high; in some soils it approaches the point of DSi saturation. Ge/Si values are lowest in these young topsoils. As with the old soils, something seems to be pumping Si into the topsoil.

At the old sites, the soils below 15 cm are in agreement with the MSF model as component 2. At the young sites, below 15 cm, Ge/Si is too high and does not agree with that found in streams. To put this another way, the lower soils of the old sites are in agreement with the MSF, but the upper parts have too much [Si] (Si enriched). At the young sites, the upper soils are in agreement, but the lower soils have abnormally high Ge/Si (Si depleted). The young soils are only in agreement with the MSF model in the upper layers.

These high [Si] and low Ge/Si findings in the upper soils at all sites are the result of phytolith entrainment. Next, the authors apply their data to form the mixing model (% component 1, % component 2), which I’ll skip.

As one may guess, the amount of Si exported to streams cannot exceed the supply in a long term fashion. The authors found an export of DSi at 150-5400 mol ha-1 yr-1 in the streams. In the upper zones of the soils, they found an export 400-9400 mol ha-1 yr-1. There seems to be a large surplus, so where is all the excess going? The authors propose that there is a rapid cycling of Si in the upper soils. Namely, any excess which is not exported to streams will probably be dissolved and taken up by plants: internal cycling.

Discussion. These data suggest that most Si in streams has passed through the internal vegetation Si cycle. This means that Si directly from mineral weathering passing into streams is only of a minor constituent. The authors go on to suggest that the phytoliths and other BSi is much more prone to dissolution than primary and secondary minerals.

Tropical humid soils are usually Si depleted, meaning that most of the Si is trapped in the vegetation. Any which is deposited in the soil is rapidly recycled. The high amount of [Si] in the upper soils acts as a buffer to toxic Al levels.

Conclusions. This is one of a series of papers which came out in 2005 detailing the biologic control of plants on Si.