1. Introduction

Our planet is experiencing climate warming at an unprecedented rate that can be attributed to increasing atmospheric greenhouse gas concentrations (IPCC 2013). Methane (CH4) is the second most abundant greenhouse gas in the atmosphere after carbon dioxide (CO2) and has a greenhouse warming potential 25 times higher than that of CO2. Earth’s atmospheric CH4 concentration has risen approximately 150% since 1750 and is largely attributed to human activities (UCAR 2012, Figure 1).

Figure 1. Plot showing CH4 mole fraction (Methane Flux) in parts per billion (ppb) over a 40 yr period (Figure from NOAA).

 

Wetlands are an important natural source of CH4 to the atmosphere (Saunois, 2020).  Methane flux is dependent on both hydrology and temperature. Hydrology is influenced by climate and land-use practices. Because CH4 flux depends on both temperature and hydrology, it is unknown if an increase in temperature only can lead to an increase in methane emissions.

1.1 Our Study Area:

 

Figure 2. Study area map of the PPR (red boundary) spanning 3 Canadian provinces and 5 U.S. states.

            The PPR, covering over an area of approximately 770,000  km2, consists of depressions created by activity from the last glaciation maximum, ending approximately 10,000 years ago (Perry et al., 2020). These depressions fill with water to create permanent to semi-permanent wetlands. This area extends into five U.S. states: Minnesota, Wisconsin, Iowa, North Dakota, and South Dakota as well as three Canadian provinces: Alberta, Saskatchewan, and Manitoba, as seen in Figure 2 The Prairie Pothole Region (PPR) is one of the largest wetlands systems in the world and is an important part of the regional and perhaps global ecosystem (Perry et al., 2020). There is critical importance in studying the PPR due to the region’s impact on geology, biology, and chemistry as well as the ecosystem services they offer (Perry et al., 2020).

1.2 Why is CH4 flux from the PPR important?

Figure 3. An aerial photograph of a section of the PPR with graphic showing CH4 exchange  (Image adapted  from USGS ( https://www.usgs.gov/media/images/prairie-pothole-region ))

 

Globally, wetlands contribute 20-40% of global atmospheric CH4 (Laanbroek 2010). Inland wetlands, such as the Prairie Pothole Region, are a known contributor to atmospheric CH4 (Bansal et al., 2016). Variations in CH4 transport to the atmosphere from the wetlands in North Dakota are not well understood at present. A better understanding of CH4 flux from regional systems like the PPR could (i) contribute towards a more complete global picture as well as (ii) enable better regional planning, especially of regional ecosystem services at a critical time of global warming-related to greenhouse gases in the atmosphere.

Our research project focused on two specific questions with regards to the methane fluxes from the PPR wetlands:

1. How has CH4 transport from the PPR wetlands to the atmosphere changed over time and space?

2. What are some of the factors impacting the CH4 transport from the PPR wetlands to the atmosphere?

While the USGS data release had 19 years of data from 199 sites, we narrowed our focus to 11 sites over an 8 year period from the ND region of the PPR to address the above questions.

2. Methods

Figure 4 shows the workflow our team followed to address our research questions. First, the data was downloaded from the USGS website. Afterwards, an initial visualization of the sites and parameters were performed, followed by narrowing down specific sites and years. Spatial and temporal variation of the selected sites were determined and possible relationships between methane flux and various variables were examined.

 

Figure 4. A flowchart of the workflow our team followed to answer our 2 proposed research questions.

2.1 Collection of Data

The dataset our team used for this project was compiled of data gathered from several studies over a span of 19-years (1997 – 2016) and downloaded from the USGS website (Science Data Catalog). The methods from each study were comparable and allowed for data to be combined into a single release. Data included measurements for: soil chemistry, greenhouse gases, topography, water chemistry, weather, and other variables including water depth, soil moisture and temperature, as seen in Figure 5.

Figure 5. Pandas DataFrame for greenhouse_gas.csv and soils.csv.

 

2.2 Initial overview and visualization of Data

Using Python, our team assessed two different files:  greenhouse_gas.csv and soils.csv by first reading in the files, as shown in Figure 6. A total of 199 different sites were recorded in the greenhouse gas.csv and 587 different sites were recorded in the soils.csv; many of which had N/A or missing values in the early years that were removed when importing in the files.

Figure 6. Python code used to import and clean missing values for data files.

The two CSV files included several variables and parameters; however, we focused mainly on the methane flux (measured in grams per m2 per hour) and the water depth (measured in cm) found in the greenhouse gas.csv datafile and the 6 trace elements (Cobalt (Co), Copper (Cu), Iron (Fe), Chromium (Cr), Manganese (Mn) and Sulfate-sulfur (SO42-), all of which are measured in parts per million in the soils.csv datafile.

2.3 Mean and Median for all years

            Initially, we investigated the mean and median CH4 flux for all 199 sites (Figure 7) plotted using the code shown in Figure 8. Based on these results, there was substantial variation across sites and it was determined that we needed to narrow down the number of sites moving forward in this study. We started to narrow down the sites by creating 4 batches of approximately 50 sites each.

Figure 7. Mean and Median graphs of CH4 flux for all 199 sites.

Figure 8. Python code used to plot the mean and median of all study sites.

2.4 Selection of Sites of Interest

The 199 initial sites spanning the PPR in the United States and Canada were narrowed down to 11 sites in North Dakota (Figure 9). The 11 focused sites allowed for a good observation of soil trace elements, which was a part of our team's 2nd research question. These 11 sites were further subset for the years of 2009-2016.

Figure 9. A full map of the PPR (left), a map of selected sites in North Dakota (center). A zoomed in map showing all 11 selected sites (right).

 

2.5 Selecting the variables of interest

The second research question our team proposed was addressed by determining any relationship between water depth, six trace elements (Co, Cu, Fe, Cr, Mn and SO42-) and CH4 transport for the 11 selected sites.  First, the mean of the trace element concentration was plotted using the code shown in Figure 10. Then, scatter plots (Figure 11) and correlation heat maps (Figure 12) were generated to determine any relationship between the selected variables and CH4 transport. Based on the results, a scatter plot was created for each year, and a heat map for 2011, to further assess the relationship between CH4 transport and water depth.

Figure 10. Python code that created the trace element concentration graph.

 

Figure 11. Water depth vs CH4 flux Python code for scatter plots.

Figure 12. Water depth vs CH4 flux Python code for the correlation heat map.

3. Results and Discussion

For research question 1 spatial and temporal variation of CH4 transport was determined. Figure 13 depicts the variation between each site with the T9 site showing the highest mean CH4 flux; however, the P8 site shows the highest median value. The sites have large variations of relatively small values that can be better compared using a log scale on the y-axis. There is a substantial difference between the mean and median, especially with T sites, which could indicate skewing of data due to an abnormal year(s). This was confirmed when CH4 flux over time was determined (Figure 14) wherein 2011 the levels were abnormally high. There are multiple factors that could be contributing to the increased flux in 2011, flooding that was occurring across North Dakota, could be one of them.

3.1 Spatial Variation

Figure 13. Bar plot showing the mean and median CH4 flux from the selected sites for 2009-2016.

3.2 Variation Over Time

 

Figure 14. Bar plot showing the yearly Methane flux for the selected sites from 2009 - 2016.

3.3 Trace Element Variability

The second research question investigated potential relationships between trace elements (Co, Cu, Fe, Cr, Mn, S-SO42-), water depth, and CH4 flux. Initially, trace element concentrations were plotted to determine any variation between sites. As seen in Figure 15, all 11 selected sites showed relatively high concentrations of iron (Fe) in green. Some elements such as cobalt and sulfate-sulfur showed some variation between sites while others did not greatly vary between sites.

Figure 15. Bar plot showing the mean trace element concentrations for each site.

3.4 Correlation

To determine if there are any relationships between the selected variables, a correlation heat map was created. The closer the value is to 1, the stronger the correlation is between the variables. Our results did not indicate any substantial correlations between methane flux and any of our selected variables (Figure 16). However, based on results from the analysis of CH4 flux over time, a closer look at water depth was warranted.

Figure 16. A correlation heatmap of methane, water depth, and trace elements.

Scatter plots of water depth vs methane flux show different patterns for different years (Figure 17). However, 2011, the year with flooding, stands out among the years. Figure 18 shows a correlation heatmap and scatter plot of water depth vs. methane flux for June, July, and August in 2011. The summer months were specifically selected to observe any potential relationship between CH4 flux and water depth while water levels were high due to flooding. Results determined that water depth shows a relatively higher correlation in 2011 compared to any other year.

Figure 17. Scatter plots showing water depth and CH4 flux for each year from 2009-2016.

 

 

Figure 18. Correlation heat map showing water depth vs methane flux for June, July, and August of 2011..

4. Future Work

From our selected sites have shown high CH4 flux variation (ex. P8 and T9) which need further attention to determine what factors cause the CH4 flux in these sites to stand out from the remaining sites. Investigating the specific factors that contributed to higher CH4 flux values during the 2011 flood could provide additional information to be used in predictive modeling. Finally, our study focused only on a few selected variables and sites. Therefore, it would be interesting to see how other variables that were not considered in this study (e.g., element concentrations in soil that are not considered here, types of wetlands) relate to CH4 flux variations.

5. Summary

CH4 transport from the PPR wetlands could be a significant contributor to the atmosphere CH4 at the regional level and beyond. There is substantial spatial and temporal flux variation. Large spatial variation of CH4 flux was observed between PPR wetlands in general and specifically between the selected sites focused on in this study. There were noticeable variations of  CH4 flux annually as well with a very high flux during 2011. All sites showed relatively high iron (Fe) concentrations. Our selected variables, in general, did not show a high correlation with CH4 flux. Water depth, particularly in 2011, showed a relatively higher correlation with CH4 flux than other years prompting further research mentioned in future works.

6. References

Bansal, S., Tangen, B., & Finocchiaro, R. (2016). Temperature and hydrology affect methane emissions from prairie pothole wetlands. Wetlands, 36(2), 371-381. https://doi.org/10.1007/s13157-016-0826-8.

IPCC (2013) Climate change 2013: the physical science basis. Contributions of Working Group I to the fifth assessment report of the Intergovernmental Panel on Climate Change. New York, NY.

Laanbroek, HJ. (2010). Methane emission from natural wetlands: interplay between emergent macrophytes and soil microbial processes. A mini-review. Annals of Botany, 105(1),141–153. https://doi.org/10.1093/aob/mcp201.

Mann, G. (2015). gmannppr [Data file]. USGS ScienceBase-Catalog. Retrieved from https://www.sciencebase.gov/catalog/item/54aeaef2e4b0cdd4a5caedf1.

Natural Resources Canada. (1999). Canadian Provinces [Data file]. National Weather Service. Retrieved from https://www.weather.gov/gis/CanadianProvinces.

Perry, C.H., Tangen, B. and Bansal, S. (2020). Great Plains. In: R.V. Pouyat, D. S.Page-Dumroese, T. Patel-Waynend, L.H. Geiser (Eds), Forest and rangeland soils of the United States under changing conditions: a comprehensive science synthesis (pp. 236-248). Springer Nature. https://link.springer.com/book/10.1007%2F978-3-030-45216-2.

Populated Places [Data file]. Natural Earth. Retrieved from https://www.naturalearthdata.com/http//www.naturalearthdata.com/download/50m/cultural/ne_50m_populated_places_simple.zip.

Saunois, M., Stavert, A. R., Poulter, B., Bousquet, P., Canadell, J. G., Jackson, R. B., ... & Zhuang, Q. (2020). The global methane budget 2000–2017. Earth System Science Data, 12(3), 1561-1623. https://doi.org/10.5194/essd-12-1561-2020.

Tangen, B., & Bansal, S. (2019). Soil properties and greenhouse gas fluxes of Prairie Pothole Region wetlands: a comprehensive data release [Data set]. U.S. Geological Survey. https://doi.org/10.5066/F7KS6QG2.

University Corporation for Atmospheric Sciences Center for Science Education. (2012). Methane. UCAR Center for Science Education. https://scied.ucar.edu/learning-zone/how-climate-works/methane.

U.S. Department of Commerce, U.S. Census Bureau, Geography Division. (2012). TIGER/Line Shapefile, 2012, nation, U.S., Current State and Equivalent National [Data file]. United States Census Bureau. http://www2.census.gov/geo/tiger/TIGER2012/STATE/tl_2012_us_state.z