Response Comments to Liska et al. (2014): Biofuels for crop residue can reduce soil carbon and increase CO2 emissions

Source: Douglas L. Karlen, USDA-ARS/NLAE • Posted: Tuesday, April 29, 2014


Response Comments to Liska et al. (2014)

Biofuels for crop residue can reduce soil carbon and increase CO2 emissions

Douglas L. Karlen, USDA-ARS/NLAE

This article makes unrealistic assumptions and uses citations out of context to reinforce the

authors’ viewpoint. For example, the first sentence citing Wilhelm et al. (2007) fails to make the

point that the authors of that publication were pointing out that the initial wind and water erosion

projections associated with the Billion Ton Study (BTS), though important, were not sufficient to

assess sustainability of the practice and that “excessive” residue removal can decrease soil

organic matter. The opening sentence would be acceptable if it stated that ‘Excessive removal …

The assumption by Liska et al. (2014) that 6 Mg/ha (2.68 tons/acre) of residue can uniformly be

removed each year is also unrealistic, even the highest producing Corn Belt states (i.e., Iowa,

Illinois, Minnesota) and for the irrigated areas in Nebraska. As stated by several other authors,

stover harvest for any use must be site specific, with “site” being best defined as sub-field units.

Several recent publications also emphasize that stover harvest guidelines must account for spatial

variability in soil resources, management practices, and weather variability. To illustrate how

unrealistic the universal 6 Mg/ha removal estimate is, I computed mean corn grain and stover

yields for the 10 Corn Belt states [Illinois, Indiana, Iowa, Michigan, Minnesota, Missouri,

Nebraska (irrigated and non-irrigated), Ohio, South Dakota, and Wisconsin] using National

Agricultural Statistics Service (NASS) data for 2003 through 2013. I then divided the assumed

removal rate (2.68 tons/acre) by the average stover production based on a harvest index (mass of

dry grain/total above ground biomass) of 0.5. Those results are:

Average NASS grain yield 153 bu/acre

Calculated total stover production 3.61 tons/acre

Amount harvested based on 6 Mg/ha removal 75%

For the highest average non-irrigated yields (Iowa, Illinois, Minnesota) and irrigated yields in

Nebraska, the average removal associated with the 6 Mg/ha assumption would be:

State % State %

Iowa 68 Minnesota 70

Illinois 70 Nebraska — irrigated 60

In contrast, the measured, 4-year average stover harvest at the POET-DSM Project Liberty site

near Emmetsburg, IA ranged from 12 to 60%, with treatments exceeding 25% being considered

excessive. Collecting 75% of the stover at or after grain harvest is also mechanically impossible

without also collecting an excessive amount of soil which fouls the conversion processes.

Harvesting 75% of all corn stover produced in the 10 Corn Belt states is unrealistic, far greater

than any projections associated with the BTS or the Revised Billion Ton Report (BT2), and

would indeed likely result in a depletion of soil organic matter.

A more realistic and balanced perspective on the potential impacts of stover harvest has recently

been published on-line and will be available in print in the 2014 volume 2 of BioEnergy

Research. That series of papers summarizes 239 site-years of corn grain and stover yields from

36 research sites and also presents estimates of the additional N-P-K removed by harvesting an

average of 3.9 or 7.2 Mg ha-1 (1.7 or 3.2 tons/acre) of stover from 28 of those sites. Another

paper in that issue uses the current studies as well as an extensive literature base to approximate

the minimum amount of crop residue that needs to be returned to sustain SOC levels. Their

results emphasize the extreme variability associated with different soils, weather patterns, and

crop growth conditions by showing that the estimated average minimum residue return rate for

35 studies was 6.38 ± 2.19 Mg stover ha-1 yr-1 (2.85 ± 0.98 tons/acre). The most important point

of that article, however, is that it refutes any notion that there is a universal minimum residue

requirement; rather it reinforces the need for field, or better yet, subfield management decisions.

Greenhouse gas (GHG) responses to stover harvest treatments were also reported. One study

summarized static chamber estimates of GHG emissions from nine corn production systems

under various crop residue and tillage management practices across the USA Corn Belt. It

showed that stover harvest generally decreased total soil CO2 and N2O emissions by -4 and -7%,

respectively, when compared to no stover removal. Decreased emissions were attributed to less

stover C and N inputs and possible microclimate differences due to changes in soil cover.

Another study showed no significant difference in N2O emission as a function of stover harvest,

but CO2 loss from the full removal plots was slightly lower than from the zero removal plots.

However, the emission difference between the two treatments was much smaller than the amount

of C removed with the stover. This implies that C was being lost from the full removal plots – a

phenomenon confirmed by rigorous soil sampling, and expected because the full-removal

treatment was considered an “excessive” rate of harvest.

Another article within the special issue of BioEnergy Research and a recent publication in

Agrociencia Uruguay provide on-site research results from the POET-DSM Project Liberty site

near Emmetsburg, IA. In the Agrociencia article, Figure 3 does show a decline in soil organic

matter (SOM) during the first four years of research at the Project Liberty site, but the decrease is

only from 4.5 to 4.0% SOM. Data from the next year (2012) which is not yet published shows a

value of 4.2% SOM. Furthermore, in 2013 a grid sampling, consisting of 131 samples for the 104

acre research site (0.79 acre cell size) showed an average SOM content of 4.4%, with sample

variation ranging from 2.4% to 6.2% SOM independent of any prior stover harvest treatments.

The SOM data as well as that for pH, buffer pH, P, K, and B levels showed substantial variation

across the site which encompasses four different Des Moines Lobe soils [Canisteo (58%),

Clarion (17%), Okoboji (17%), and Nicollet (8%)]. This suggests the initial SOM trend was

influenced more by sample variation than stover harvest. Furthermore, in both publications, it is

concluded that tillage rather than moderate stover harvest that is affecting SOM.

The bottom line is that stover harvest for any use must be site specific and strive to balance all

factors affecting soil carbon and all other factors that influence soil health and sustainability.