Investigating redox drivers across regions

What are hydric soils?


Hydric soils form when anaerobic conditions develop under sustained saturation (via flooding or ponding) during the growing season.


Description of the image

Hydric soils occur across the United States. From coastal Oregon soils to temperate wetlands in Minnesota to southern pine forests, these biologically significant soils develop across a variety of climates, topographies, vegetation, and parent materials.


What drives oxygen depletion?

STUDY AREA

Study areas were selected using Web Soil Survey, SoilWeb, and NASIS to find geographically distinct areas that contained hydric soils.

For this project, we will explore hydric soils in six geographically contrasting wildland sites:

Site State Soil Survey Area MLRA
Siuslaw National Forest Oregon OR638 4A
Battleground State Park Minnesota MN021 93A
Snake River Valley Wyoming WY666 43B
Targhee National Forest Idaho ID758 43B
De Soto National Forest Mississippi MS131 133C
Yellow River Wildlife Management Area Florida FL113 133C


OBJECTIVES AND HYPOTHESES


Overarching Objective: Investigate drivers of redox conditions in hydric soils.


Hypotheses


RESULTS AND DISCUSSION

First, we need to compare the amount of hydric soils at each site. Here, the term site is used to reference each individual study area which represents a selected soil survey area. See site table for those individual soil survey areas. Relevant forests included within the soil survey areas are used in site terminology, as they are the locations I was targeting when selecting the soil survey areas.

To do this, we subset our dataset to select only soils that are some percent hydric (greater than zero).

The sites all have different amounts of pedons that are > 0 % hydric. From most soil with hydric soils to the least:



Characteristics of redox-affected soils

Now, it is immediately apparent that some amount of this separation in amount of hydric soils across sites is accounted for by climate.


But, before diving too much further into how these sites are different, let’s look at what common drivers in redox may be occurring across all of these sites.

Subset soils from the extracted NASIS pedons that include one of the following

  • reduced matrix
  • redox concentrations
  • redox depletions with a chroma 2 or less

Of the remaining pedons across all the sites, plot the clay, pH, and gravel

Individual Soil Profiles

Identifying Diagnostic Features that may indicate redox conditions

Here we assess pedons from our four lowest hydric pedon count site areas. We are looking for any pedons that contain a diagnostic feature that we associate with redox conditions.


pedons


Our analysis only yields pedons from Mississippi, Oregon, and Wyoming. No soils from the Florida site met our criteria, likely because there is an usually high proportion of soils that have NA data.

In the table below, soils are listed in the same order as the figure (left to right).

Soil Series SSA Aquic Conditions Reduced Matrix Redox Conc. Redox Depl.
Saucier MS131 X
Coquille OR041 X X
Wilsonville WY039 X
Chitwood OR041 X
Chitwood OR041 X
Coquille OR041 X X X X

Please note that the two Oregon soil series have two contrasting pedons each, which is the reason you must refer to the figure to understand how they are listed in the table. In the case of Chitwood, the soil series are identical, and I’m not sure what differentiates them


The Saucier soil is an interesting one to use to test our hypotheses (keeping in mind that this is a limited analysis and may over interpret the data).

While aquic conditions occur, likely as a combination of climate and a shallow water table associated with plinthite, there appears to be little organic content (as far as we can tell from this data). That could indicate that while saturation is present, oxygen consumption is limited due to constraints in substrate availability.

I would expect the Oregon soils to have more organic carbon, based on the climate. But we can’t know for sure just by looking at these profiles.


So…

I guess we better go look at some KSSL data!

Oh wow!

It looks like Saucier is, indeed, very low in organic carbon content throughout the profile.

We can compare with Oregon’s Coquille which does indeed have quite a bit more organic carbon.


While we’re here, let’s take a look at clay too!

Definitely a lot of clay in those Missippi soils. Also quite a bit in Chitwood, the other Oregon soil. That makes sense, there are definitely some ultisols up there!



Let’s do another subset of our data!


Subsetting using a histogram

There are 765 pedons that are greater than zero percent hydric.

nrow(mu_dat)
## [1] 765

Of those 765 pedons that have some amount of hydric soil, the majority are less than 25. The next highest frequency of soils that are some percent hydric occurs at the opposite end of the spectrum (over 75 % hydric).

## `stat_bin()` using `bins = 30`. Pick better value `binwidth`.

In order to do a more targeted analysis, we will subset the dataset to focus on soils most likely to be hydric. To create this subset, we will select only soils that are > 75 percent hydric.

Individual Soil Profiles

Look at these gorgeous soil profiles. Notably, Atmore has plinthite, did our other MS soil: Saucier.

Rains is our Florida soil, and one of the most reduced. This aligns with what I would expect from a swampy Florida soil! Those NAs in our earlier dataset may have done us a big disservice in finding a cool reduced Florida soil!

Of these soils, Rifle, Mooseflat, and Brenner seem likely to have the most organic matter.

Climate

Interesting. Rifle and Mooseflat (MN and ID, respectively) are some of the coldest soils (along with WY Newfork). So, does this mean we’re accumulating organic matter? Are we limiting microbial metabolism in colder temperatures or providing more substrate to decompose and consume oxygen? Rifle and Mooseflat are the only soils with an organic horizon … or six organic horizons, in Rifle’s case.

It’s a question I run into a lot in my research, and I still don’t know how to account for temperature in the microbial oxygen consumption and substrate generation balance.

However, it’s worth noting that ID and MN had the most > 75 % hydric soils of all site areas.

Monthly Climate

Look at the Mediterranean climate represented by Brenner along the Oregon Coast!

The timing of precipitation may matter as well, which is one reason I love these plots. In Atmore and Brenner, the majority of precipitation falls in the summer: the hottest months. That’s also when Potential ET is highest. Is that why FL and MS were on the lower end of > 75 % hydric soils? Does that, along with lack of substrate availability, limit redox fluctuations?

Hillslope and Geomorphology

No surprise that toeslopes are dominating these > 75 % hydric soils!

Rains is interesting, with the summit position’s representation. And yet, that soil was quite gleyed! However, if we take a look at the map, we can see that Rains is a widespread soil… so we’re representing quite a few pedons here!

CONCLUSIONS

As with all studies, this one raises more questions than it answers.

My conclusion is that the factors that drive redox conditions vary geographically!

If only I could install redox probes in every soil in the US! We would see some very interesting seasonal dynamics.

For now, these are the key concepts that I plan to think about more:

STAT 2020 Addition

## 'data.frame':    80 obs. of  12 variables:
##  $ series     : chr  "ATMORE" "ATMORE" "ATMORE" "ATMORE" ...
##  $ climate_var: chr  "Effective Precipitation (mm)" "Mean Annual Air Temperature (degrees C)" "Elevation (m)" "Fraction of Annual PPT as Rain" ...
##  $ minimum    : num  434 18 0 100 1384 ...
##  $ q01        : num  521.3 18.2 2 100 1481.9 ...
##  $ q05        : num  542.8 18.3 4 100 1505 ...
##  $ q25        : num  575.5 18.5 19 100 1536 ...
##  $ q50        : num  607.1 18.7 40 100 1583 ...
##  $ q75        : num  641.9 19.4 59 100 1653 ...
##  $ q95        : num  672.9 19.6 83 100 1706 ...
##  $ q99        : num  702.3 19.6 117.2 100 1750.1 ...
##  $ maximum    : num  751.4 19.7 155 100 1791 ...
##  $ n          : int  4890 4890 4890 4890 4890 4890 4890 4890 278 278 ...
##           Design Freeze Index (degrees C) Effective Precipitation (mm)
## ATMORE                                 26                    607.09216
## BRENNER                                27                   1463.27740
## CHITWOOD                               27                   1584.73230
## COQUILLE                               17                   1248.32629
## MOOSEFLAT                            1143                    -15.84467
## NEWFORK                               988                    -34.68781
##           Elevation (m) Fraction of Annual PPT as Rain Frost-Free Days
## ATMORE             40.0                            100             254
## BRENNER            35.5                             99             246
## CHITWOOD           36.0                             99             242
## COQUILLE           10.0                             99             265
## MOOSEFLAT        1962.0                             78              75
## NEWFORK          2435.0                             77              87
##           Growing Degree Days (degrees C)
## ATMORE                             3529.0
## BRENNER                            1161.5
## CHITWOOD                           1141.0
## COQUILLE                           1142.5
## MOOSEFLAT                           861.5
## NEWFORK                             995.0
##           Mean Annual Air Temperature (degrees C)
## ATMORE                                  18.735134
## BRENNER                                 11.018567
## CHITWOOD                                10.980069
## COQUILLE                                10.993950
## MOOSEFLAT                                2.845994
## NEWFORK                                  3.674289
##           Mean Annual Precipitation (mm)
## ATMORE                            1583.0
## BRENNER                           2129.5
## CHITWOOD                          2248.0
## COQUILLE                          1891.0
## MOOSEFLAT                          433.0
## NEWFORK                            438.0
## Dissimilarities :
##             ATMORE BRENNER CHITWOOD COQUILLE MOOSEFLAT NEWFORK RAINS RIFLE
## BRENNER       3.35                                                        
## CHITWOOD      3.56    0.32                                                
## COQUILLE      3.08    0.63     0.94                                       
## MOOSEFLAT     6.22    6.01     6.21     5.75                              
## NEWFORK       6.23    6.13     6.32     5.86      0.66                    
## RAINS         1.06    3.64     3.90     3.26      5.53    5.56            
## RIFLE         5.26    5.01     5.21     4.72      2.47    2.98  4.65      
## SAUCIER       0.31    3.48     3.67     3.23      6.47    6.47  1.31  5.51
## WILSONVILLE   6.54    6.25     6.41     6.04      0.98    1.16  5.94  2.76
##             SAUCIER
## BRENNER            
## CHITWOOD           
## COQUILLE           
## MOOSEFLAT          
## NEWFORK            
## RAINS              
## RIFLE              
## SAUCIER            
## WILSONVILLE    6.77
## 
## Metric :  euclidean 
## Number of objects : 10

## Initial stress        : 0.01980
## stress after  10 iters: 0.00453, magic = 0.068
## stress after  20 iters: 0.00219, magic = 0.500
## stress after  30 iters: 0.00215, magic = 0.500

## Run 0 stress 0.001634267 
## Run 1 stress 0.001691795 
## ... Procrustes: rmse 0.02613558  max resid 0.03819425 
## Run 2 stress 0.001377214 
## ... New best solution
## ... Procrustes: rmse 0.02259575  max resid 0.03500646 
## Run 3 stress 0.0002928714 
## ... New best solution
## ... Procrustes: rmse 0.03756081  max resid 0.05819401 
## Run 4 stress 0.0009580848 
## Run 5 stress 0.002093324 
## Run 6 stress 0.002059344 
## Run 7 stress 0.001460113 
## Run 8 stress 0.002818151 
## Run 9 stress 9.735401e-05 
## ... New best solution
## ... Procrustes: rmse 0.02134815  max resid 0.03869611 
## Run 10 stress 0.143587 
## Run 11 stress 0.001732311 
## Run 12 stress 0.002576199 
## Run 13 stress 0.00205349 
## Run 14 stress 0.002140367 
## Run 15 stress 0.00145642 
## Run 16 stress 0.143587 
## Run 17 stress 0.000672715 
## Run 18 stress 0.002565471 
## Run 19 stress 0.001116948 
## Run 20 stress 9.929305e-05 
## ... Procrustes: rmse 0.0169244  max resid 0.0324833 
## *** Best solution was not repeated -- monoMDS stopping criteria:
##     16: no. of iterations >= maxit
##      2: stress < smin
##      2: scale factor of the gradient < sfgrmin
## Warning in metaMDS(sp4.scaled[, -1], distance = "gower", autotransform = FALSE,
## : stress is (nearly) zero: you may have insufficient data
## species scores not available

## species scores not available

Interesting! Oregon coast and Gulf coast had similar climates while Minnesota and Wyoming/Idaho had similar climates. Those similarities did not translate to similar redox conditions!

It’s interesting to consider which climate parameters are most similar across these dendrogram specified groupings (design freeze index, elevation, MAP) and which are not similar (effective precip, growing degree days, MAAT). What do these similarities and differences tell us about climate controls on redox conditions?

REFERENCES

Herndon, E., A. AlBashaireh, D. Singer, T. Roy Chowdhury, B. Gu, et al. 2017. Influence of iron redox cycling on organo-mineral associations in Arctic tundra soil. Geochimica et Cosmochimica Acta 207: 210–231. doi: 10.1016/j.gca.2017.02.034.

Herndon, E.M., L. Kinsman‐Costello, K.A. Duroe, J. Mills, E.S. Kane, et al. 2019. Iron (Oxyhydr)Oxides Serve as Phosphate Traps in Tundra and Boreal Peat Soils. J. Geophys. Res. Biogeosci. 124(2): 227–246. doi: 10.1029/2018JG004776.

Herndon, E., L. Kinsman‐Costello, and S. Godsey. 2020. Biogeochemical Cycling of Redox‐Sensitive Elements in Permafrost‐Affected Ecosystems. In: Dontsova, K., Balogh‐Brunstad, Z., and Le Roux, G., editors, Geophysical Monograph Series. 1st ed. Wiley. p. 245–265

Herndon, E.M., B.F. Mann, T. Roy Chowdhury, Z. Yang, S.D. Wullschleger, et al. 2015. Pathways of anaerobic organic matter decomposition in tundra soils from Barrow, Alaska: BIOGEOCHEMISTRY OF ANOXIC ARCTIC TUNDRA. J. Geophys. Res. Biogeosci. 120(11): 2345–2359. doi: 10.1002/2015JG003147.

Rooney, E.C., E. VanderJeugdt, S. Avasarala, I. Miah, M.J. Berens, et al. 2024. Decoupling of redox processes from soil saturation in Arctic tundra. Communications Earth & Environment 5(1): 746. doi: 10.1038/s43247-024-01927-1.

Fiedler, S., and M. Sommer. 2000. Methane emissions, groundwater levels and redox potentials of common wetland soils in a temperate‐humid climate. Global Biogeochemical Cycles 14(4): 1081–1093. doi: 10.1029/1999GB001255.

Fiedler, S., and M. Sommer. 2004. Water and Redox Conditions in Wetland Soils—Their Influence on Pedogenic Oxides and Morphology. Soil Science Society of America Journal 68(1): 326–335. doi: 10.2136/sssaj2004.3260.

Lacroix, E.M., M. Aeppli, K. Boye, E. Brodie, S. Fendorf, et al. 2023. Consider the Anoxic Microsite: Acknowledging and Appreciating Spatiotemporal Redox Heterogeneity in Soils and Sediments. ACS Earth Space Chem. doi: 10.1021/acsearthspacechem.3c00032.

LaCroix, R.E., M.M. Tfaily, M. McCreight, M.E. Jones, L. Spokas, et al. 2019. Shifting mineral and redox controls on carbon cycling in seasonally flooded mineral soils. Biogeosciences 16(13): 2573–2589. doi: 10.5194/bg-16-2573-2019.