Next Article in Journal
Multiscale Remote Sensing Data Integration for Gully Erosion Monitoring in Southern Brazil: Case Study
Previous Article in Journal
Image Classification of Chicken Breed and Gender Using Deep Learning
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
AgriEngineering
Volume 7
Issue 7
10.3390/agriengineering7070213
agriengineering-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Evaluation of Greenhouse Gas-Flux-Determination Models and Calculation in Southeast Arkansas Cotton Production

by
Cassandra Seuferling
,
Kristofor Brye
,
Diego Della Lunga
*,
Jonathan Brye
,
Michael Daniels
,
Lisa Wood
and
Kelsey Greub
Department of Crop, Soil, and Environmental Science, University of Arkansas, Fayetteville, AR 72701, USA
*
Author to whom correspondence should be addressed.
AgriEngineering 2025, 7(7), 213; https://doi.org/10.3390/agriengineering7070213
Submission received: 2 June 2025 / Revised: 22 June 2025 / Accepted: 27 June 2025 / Published: 2 July 2025
Browse Figures
Review Reports Versions Notes

Abstract

Greenhouse gas (GHG) emissions evaluations from agroecosystems are critical, particularly as technology improves. Consistent GHG measurement methods are essential to the evaluation of GHG emissions. The objective of the study was to evaluate potential differences in gas-flux-determination (GFD) options and carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O) fluxes and growing-season-long emissions estimates from furrow-irrigated cotton (Gossypium hirsutum) in southeast Arkansas. Four GFD methods were evaluated [i.e., linear (L) or exponential (E) regression models, with negative fluxes (WNF) included in the dataset or replacing negative fluxes (RNF)] over the 2024 growing season using a LI-COR field-portable chamber and gas analyzers. Exponential regression models were influenced by abnormal CO2 and N2O gas concentration data points, indicating the use of caution with E models. Season-long CH4 emissions differed (p < 0.05) between the WNF (−0.51 kg ha−1 season−1 for L and−0.54 kg ha−1 season−1 for E) and RNF (0.01 kg ha−1 season−1 for L and E) GFD methods, concluding that RNF options over-estimate CH4 emissions. Gas concentration measurements following chamber closure should remain under 300 s, with one concentration measurement obtained per second. The choice of GFD method needs careful consideration to result in accurate GHG fluxes and season-long emission estimates.

1. Introduction

Climate change has become a recent global concern, particularly regarding greenhouse gas (GHG) emissions of carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O). The Intergovernmental Panel on Climate Change (IPCC) reported greater GHG concentrations presently than in the past 800,000 years due to human activity [1]. The IPCC also reported that, from 2007 to 2016, agriculture, forestry, and other land uses contributed 23% to the net global anthropogenic GHG emissions [2]. Consequently, the focus of GHG emission reduction and mitigation has turned towards agricultural soil and cropland management. The growing scientific effort to characterize GHG emissions in agricultural systems required the adoption of reliable standardized sampling protocols along with shared analytical approaches.
Techniques of GHG measurements and emissions estimates have changed substantially over time. Techniques range from static, closed-chamber approaches requiring manual gas sample collection and transport, to laboratory-based gas chromatography equipment [3,4,5,6,7,8], to dynamic, portable, closed chambers [9,10,11] and flow-through analyzers, allowing for in-field measurements with an infrared gas analyzer (IRGA). In particular, LI-COR Environmental (Lincoln, NE, USA) manufactures and markets portable, closed chambers and IRGAs for quick, in-field measurements of CO2, CH4, and N2O, with one measurement of each gas every second [12,13]. However, several factors of data collection and post-processing are present and vary among studies [14,15,16,17].
The measurement duration or amount of time the chamber is closed (i.e., observation length) and has yet to be evaluated under field conditions for an extended period of time in various agronomic settings. Gas concentration measurements may be impacted by gas build-up in the chamber headspace, potentially reversing the gas-flow direction from out of the soil back into the soil, resulting in decreased measured gas concentrations after some time [18]. At the initial closing of the chamber, a gas disturbance occurs. Thus, a short period of steady-state mixing must be established before gas concentration data can be considered representative, referred to as a deadband, and is dealt with in the post-processing of gas concentration data [18]. The closed chamber creates an artificial environment between the soil and chamber where air temperature commonly increases every minute, affecting the concentration of the released gases from the soil according to the ideal gas law [19]. As a consequence, a long observation length when GHG production is relatively small or negligible can result in a negative flux that is not representative of the biogeochemical processes under evaluation at that moment [19].
Similar to measurement duration and deadband period, when calculating GHG fluxes from GHG concentration data over time, either a linear or non-linear (e.g., exponential) regression model may be used, depending on the shape of the resulting concentration-over-time relationship and on the nature of the biogeochemical process for a specific gas. However, discrepancies among GHG flux results from linear and non-linear models have been observed in closed-chamber systems [20,21,22,23,24,25]. Flux calculations after measurement can result in GHG fluxes that are positive or negative. Greenhouse gas studies either include the negative flux value in the overall evaluation of GHG fluxes and season-long emissions [26,27,28] or reassign the negative value to a small, yet positive, flux value [29], under the assumption that a negative flux is the result of chamber error or chamber artificial environment or to solely evaluate GHG emissions rather than the uptake of gases into the soil [30,31,32,33,34]. Differences in measurement duration and the usage of various regression models, with or without negative fluxes, among studies may result in data discrepancies, methodological variations, and inaccurate reporting for emissions inventories.
The presence of regional and global climate change indicates a need for GHG research, particularly with agricultural practices thought to have a positive impact on climate, and the evaluation of methods to mitigate GHG emissions in major cropping systems will contribute to efforts in reducing the potential detrimental impacts of climate change. As a major, global, non-food commodity, cotton (Gossypium hirsutum) is currently gaining interest for sustainable production to enhance its marketability, which includes minimizing the GHG emissions and global warming potential (GWP). In 2023, the United States (US) contributed 11% to global cotton production [35] and is the world’s largest cotton exporter [36]. Arkansas ranked fourth in nationwide cotton production in 2023, as 6% of the nationwide production [35], primarily grown on the eastern third of Arkansas in the Lower Mississippi River Valley (LMRV) [37]. In Arkansas, the economic importance of cotton requires further study of GHG production in climate-smart cotton production. However, no studies have been conducted on GHG emissions from cotton production systems in the LMRV of eastern Arkansas. Furthermore, to date and to the authors’ knowledge, there are no comprehensive national (i.e., US-based studies) or global summaries of GHG emissions from cotton.
Therefore, the overall objective of this field study was to evaluate potential differences in GHG measurement procedures, flux-determination options, and CO2, CH4, and N2O fluxes and emissions estimates from furrow-irrigated cotton in southeast Arkansas. Two specific sub-objectives for this study were to (i) evaluate four gas-flux determination methods on season-long fluxes and emissions using linear (L) or exponential (E) regression models, with negative fluxes (WNF) included in the dataset or replacing negative fluxes (RNF) with a small, positive value, and (ii) evaluate measurement duration [i.e., length of time the chamber is closed and measuring gas concentrations; 120, 300, 600, 900, 1200, 1500, 1800, 2100, 2400, and 2700 s, and a business-as-usual (BAU) option] on resulting GHG fluxes.
For sub-objective i, it was hypothesized that the linear-regression flux-determination method would result in a lower (i.e., under-estimation) CO2, CH4, and N2O flux, regardless of WNF or RNF, compared to the exponential-regression method based on past study results [22,25,38] and the assumption that gases are expelled from the soil in a non-linear manner [39]. It was also hypothesized that the WNF option would result in lower GHG fluxes than the RNF option, regardless of the regression model used, due to the potential for negative fluxes to be significantly different from zero. For sub-objective ii, it was hypothesized that CO2, CH4, and N2O fluxes will not differ among different measurement durations based on manufacturer recommendations and the assumption of gases reaching a steady state within 45 s from the start of a measurement and remain steady within the chamber headspace for the 120 and 320 s measurement durations for CO2/CH4 and N2O, respectively [40,41]. The novelty of this study lies in the simultaneous in-field measurement of CO2, CH4, and N2O over time using state-of-the-art, field-portable analyzers to generate high-quality datasets to evaluate data-analysis options.

2. Materials and Methods

2.1. Site Description

Field research was conducted in a private landowner’s field near Tillar in Desha County, Arkansas (33.815° N, −91.34° W). The field and surrounding area are located within the Arkansas River Alluvium Major Land Resource Area, consisting of agroecosystems and forested wetlands with a variety of soil textures [42]. This study was conducted on a Hebert silt loam (fine-silty, mixed, active, thermic Aeric Epiaqualfs), which is a deep, somewhat poorly drained, and moderately slowly permeable soil formed in a silty alluvium [43,44].
The study area comprised 9 raised beds in an approximately 49 m long (161 ft) and 5.6 m wide (18.4 ft) area and was located approximately 46 m down slope from the north edge of the field (Figure 1). The current study comprised four plots within the study area arranged in a complete randomized design, and each plot was considered an independent replicate.
Each single experimental unit’s total area was a rectangular area of ~2.9 m2 that spanned the center of one furrow to the center of the adjacent furrow containing one raised bed (i.e., 0.7 m in width) and 1.5 m in both directions from a single, 20 cm diameter GHG sampling base collar (described below) installed in the middle of the raised bed.
The 30-year (1991–2020) monthly average air temperature was 17.8 °C, and the 30-year (1991–2020) average annual precipitation was 130 cm in the study region [45]. The maximum mean monthly air temperature occurs in July and the largest mean monthly rainfall occurs in April [45].

2.2. Treatments and Experimental Design

For objective i, four specific gas-flux-determination (GFD) methods were evaluated (i.e., LWNF, LRNF, EWNF, and ERNF) in a completely random design with time (i.e., weekly measurements) as a repeated measure. For objective ii, linear and exponential regression models, with negative fluxes included, were fit to 11 specific gas-flux-determination durations for evaluation, including 120, 300, 600, 900, 1200, 1500, 1800, 2100, 2400, and 2700 s for all three gases and a business-as-usual BAU option, where CO2/CH4 and N2O flux determinations were set to the subsequent 120 and 320 s, respectively, after a 45 s deadband duration for all three gases, in a completely random design.

2.3. Field and Plot Management

Since 2022, the field has been planted with only cotton, minimally tilled, and managed with furrow irrigation as needed throughout the growing season with well water. Minimal tillage included one pass with a roller bedder and one pass with a bed conditioner without cover crop (CC) planted. The cotton variety Deltapine (Scott, MS, USA) DPL 2127B3XF was planted on 25 April 2024 [day of year (DOY) 116] at 168,127 kg seed ha−1 at ~2.5 cm deep with ~15 cm in-row spacing on raised beds that were 65 to 70 cm wide (i.e., bed center to adjacent bed center) and ~9 cm tall. Broadcast applications of 112 kg ha−1 diammonium phosphate and ammonium sulfate, 224 kg ha−1 muriate of potash, and 1.1 kg ha−1 boron salt occurred on 10 June 2024 (DOY 162) without incorporation. An airplane-applied tank mixture of 584 mL ha−1 of boric acid (BoronPlus, DeltAg, Greenville, MS, USA) and 876 mL ha−1 prosulfuron (Peak, Syngenta, Basil, Switzerland) were applied to the study area on 20 June 2024 (DOY 172). Diagonal knife injection into the planted beds at a 5 cm depth of 87 L urea ammonium nitrate ha−1 occurred on 25 June 2024 (DOY 177). Another tank mixture of 1.02 L ha−1 of prosulfuron, 110 mL ha−1 of sulfoxaflor (Transform, Corteva, Indianapolis, IN, USA), and 584 mL ha−1 novaluron (Diamond, ADAMA Insecticide, Ashdod, Israel) pesticides, with 584 mL ha−1 of boric acid, were applied on 18 July 2024 (DOY 200) to the study area.

2.4. Gas, Soil Moisture, and EC Measurements

Gas sampling locations consisted of a thick-walled, 11.5 cm tall, 20-cm diameter, polyvinyl chloride (PVC) base collar, with a beveled bottom. Manual base-collar installation occurred after cotton planting on 25 April 2024 (DOY 116) by pounding the PVC collar to a depth of ~9 cm. After base collar installation, the distance from the soil surface to the top of the collar was measured at four points within the collar to represent the collar offset for subsequent GHG flux determinations.
All GHG concentration measurements were conducted using a field-portable LI-COR Smart Chamber (model LI-8200-01S, LI-COR, Lincoln, NE, USA, 4244.1 cm3) connected to field-portable LI-COR CO2/CH4 (model LI-7810, LI-COR, Lincoln, NE, USA) and N2O (model LI-7820, LI-COR, Lincoln, NE, USA) analyzers using optical feedback-cavity enhanced absorption spectroscopy. The protocols developed for gas-flux determinations in the current study were adapted from the standard operating procedure suggested by LI-COR for in-field measurement [40,41]. A combined soil moisture/electrical conductivity (EC) probe (HydraProbe SDI-12, Stevens Water Monitoring Systems, Inc., Portland, OR, USA) was also connected to the Smart Chamber for simultaneous measurement during GHG concentration measurements. The Smart Chamber was placed atop the base collar, without plants inside, for each measurement and the ~5 cm long soil moisture/EC probe was inserted into the raised bed ~0.3 m up-slope from the base collar. Before each measurement, any plant matter and/or weeds were removed from the base collars such that only soil gases were measured. The Smart Chamber was programmed to automatically close to create an air-tight system and split the air flow to each gas analyzer through 2.15 m of tubing for simultaneous GHG concentration measurements. Gas concentrations were measured and recorded every second for the total duration the Smart Chamber was closed on the base collar, which was 320 s. After each measurement was completed, the Smart Chamber was programmed to automatically open and purge all tubing and both analyzers by pumping ambient air through it for 45 s to ensure the removal of any remaining residual gas from the previous measurement.
For Objective i, gas sampling events occurred weekly from 8 May 2024 (DOY 129) to the end of the growing season, with the last gas sampling on 10 September 2024 (DOY 254) for a total of 19 sampling dates. The desired sampling time of day was between 0700 and 1000 h to capture the time period encompassing the daily average air temperature, which has been determined to be the optimal daily sampling time period [29]. Deviations from the desired sampling time only occurred when weather conditions did not permit morning sampling, when gas sampling occurred between 1700 and 2000 h to capture the second window for the daily average air temperature [29].
Two other separate sampling events occurred on 2 July and 6 August 2024 (DOY 184 and 219, respectively) between 1400 and 1800 h for data collection for objective ii. Measurements followed the same procedures as for objective i, except that gas concentrations were measured continuously for 2700 s (45 min), assuming that 2700 s was a sufficient duration to identify potential differences in calculated GHG fluxes. Due to a collar labeling error, only three replications of the four plots were measured on 2 July, while all four replications were measured on 6 August 2024.

2.5. Data Processing

Once sampling for a date was completed, data were downloaded from the Smart Chamber and analyzers and imported into the Soil Flux Pro software (version 5.3) to calculate the dry flux value for each replicate measurement. Gas fluxes were calculated as the gas concentration change per unit area per unit time (i.e., μmol m−2 s1 and nmol m−2 s−1 for CO2/CH4 and N2O, respectively). For objective i, in Soil Flux Pro, the collar offset was input, the deadband was set to 45 s, and the measurement durations for CO2/CH4 and N2O flux determinations were set to the subsequent 120 and 320 s, respectively, after the 45 s deadband duration, which was considered the BAU option. Default values in the Soil Flux Pro software were used for all other possible parameters [41]. Fluxes were calculated with the four GFD methods: LWNF, LRNF, EWNF, and ERNF, where RNF represented reassignment of any originally calculated negative flux to the value of 1.0 × 10−6 μmol m−2 s−1 or nmol m−2 s−1, such that all fluxes associated with the RNF option were positive. The small positive value represented the positive detection limit of the analyzers as indicated by LI-COR protocols [41].
For objective ii, the Soil Flux Pro software was used to determine each gas’s flux using the linear and exponential regression models, with negative fluxes included, for each of the 11 flux-determination durations (i.e., 120, 300, 600, 900, 1200, 1500, 1800, 2100, 2400, and 2700 s and BAU) following the 45 s deadband.

2.6. Season-Long Emissions and GWP Determinations

For objective i only, CO2, CH4, and N2O fluxes for each individual collar and GFD method (i.e., LWNF, LRNF, EWNF, and ERNF) were linearly interpolated between consecutive measurement dates and summed to determine season-long emissions. Global warming potential was also calculated for each individual collar and GFD method as the sum of CO2 equivalents from each gas’s season-long emissions using the 100 yr conversion factors of 265 for N2O and 28 for CH4 [2,46].

2.7. Statistical Analyses

Based on a completely random design, for objective i, to isolate effects of specific treatments (i.e., linear and exponential separately within RNF only and WNF only, WNF and RNF separately within linear only and exponential only), four separate two-factor analyses of variance (ANOVAs) were conducted to evaluate the effects of the treatment pairs, time (i.e., weekly measurements) and their interaction on CO2, CH4, and N2O fluxes. A one-factor ANOVA was conducted to evaluate the effects of the GFD method (i.e., LWNF, EWNF, LRNF, and ERNF) on season-long CO2, CH4, and N2O emissions and GWP.
For objective ii, a one-factor ANOVA was conducted to evaluate the effects of measurement duration (i.e., 120, 300, 600, 900, 1200, 1500, 1800, 2100, 2400, and 2700 s and BAU) on CO2, CH4, and N2O fluxes separately for both LWNF and EWNF flux datasets and separately for each of the two individual measurement dates and for both measurement dates combined. In addition, a one-factor ANOVA was conducted to evaluate the effects of the GFD option (i.e., LWNF and EWNF) on CO2, CH4, and N2O fluxes separately for each measurement duration (i.e., 120, 300, 600, 900, 1200, 1500, 1800, 2100, 2400, and 2700 s and the BAU option) separately for each of the two individual measurement dates and for both measurement dates combined. Only the LWNF and EWNF flux determination options were used for evaluating measurement duration to avoid skewing the data by reassigning negative flux values.
All statistical analyses were conducted in R Studio (version 4.3.2, R Foundation for Statistical Computing, Vienna, Austria). For all datasets, normality was checked, and a normal (Gaussian) or gamma distribution for response variables was used accordingly. Datasets were tested for homogeneity of variance. As a result, all GFD model comparisons (i.e., LRNF and ERNF, LWNF and EWNF, LWNF and LRNF, and EWNF and ERNF) were characterized as having homogeneity of variance. For CH4 and N2O, all GFD model comparisons were a normal distribution, except for GFD model comparisons LRNF and ERNF. For CO2, all GFD model comparisons were evaluated with a gamma distribution except LWNF and LRNF, evaluated as a normal distribution. Significance was judged at p ≤ 0.05, and, when necessary, means were separated by least significant difference or Tukey’s Honest Significant Difference.

3. Results

3.1. Evaluation of Gas-Flux-Determination Methods

3.1.1. CO2 Fluxes and Season-Long Emissions

Greenhouse gas fluxes during the 2024 growing season in a common furrow-irrigated cotton production system in southeast Arkansas were collected to evaluate the various GFD method datasets. Carbon dioxide fluxes differed between linear and exponential regression models over time (p < 0.05) separately within the RNF and WNF datasets (Table 1; Figure 2). In addition, CO2 fluxes did not differ (p > 0.05) between WNF and RNF methods when linear and exponential regression models were compared separately, but, averaged across WNF and RNF methods, CO2 fluxes differed (p < 0.01) over time (Table 1; Figure 2).
Since there were no negative CO2 fluxes in the final datasets, as the CO2 concentrations did not decrease over time for any measurement, the results for the linear and exponential comparison within the RNF and WNF methods were identical. Thus, only the linear and exponential comparison from the RNF method was described. For the RNF method only, CO2 fluxes were similar between the linear and exponential regression models on every measurement date, except for DOY 238, when the CO2 flux from the exponential model (2105 mg m−2 h−1) was greater than that for the linear model (692 mg m−2 h−1; Figure 2). On DOY 238, the CO2 concentration variability over time among collar replicates resulted in a larger calculated exponential flux compared to the linear model (Figure 2). The constant slope of the linear regression over time is often characterized by wider confidence intervals that can better capture the trend over time without being impacted by influential values [22,25], like the exponential model was in this particular case.
Considering only the pattern over time, initial (i.e., DOY 129) CO2 fluxes started relatively large, followed by a general decreasing trend up to DOY 212 and a decreasing trend up to DOY 220, followed by a more stable period through the end of the growing season. The numeric peak CO2 flux for the linear GFD model (1243 mg m−2 h−1) occurred on DOY 212, while the numeric peak CO2 flux for the exponential GFD model (2105 mg m−2 h−1) occurred on DOY 238. A general increase in CO2 fluxes through the mid-growing season, followed by a decrease in CO2 fluxes towards the end of the growing season, is common among GHG studies in upland agricultural systems, including soybean (Glycine max), corn (Zea mays), and cotton [47,48,49].

3.1.2. CH4 Fluxes and Season-Long Emissions

In contrast to CO2, CH4 fluxes from the RNF GFD models did not differ between linear and exponential models (p > 0.05) but varied over time (p < 0.01; Table 1). For 18 measurement dates, the CH4 flux averaged between linear and exponential RNF models was essentially zero (5.8 × 10−8 mg m−2 h−1), except for DOY 238 (Figure 3). The CH4 flux for DOY 238 was the only positive flux (0.006 mg m−2 h−1) for the entire growing season and differed from all other CH4 fluxes in the RNF dataset (Figure 3) but was not a formal outlier.
In contrast to the RNF dataset, CH4 fluxes from the WNF dataset did not differ (p > 0.05) between linear and exponential GFD models and did not vary over time (p > 0.05; Table 1). In contrast to CO2, but similar to that hypothesized, CH4 fluxes differed between RNF and WNF GFD methods for both the linear (p < 0.01) and exponential (p < 0.01) models but did not differ over time (i.e., measurement dates; p > 0.05; Table 1). For the linear model, averaged across measurement dates, the mean CH4 flux was greater from the RNF (<0.001 mg m−2 h−1) than the WNF method (−0.017 mg m−2 h−1). For the exponential model, averaged across measurement dates, the mean CH4 flux was greater from RNF (<0.001 mg m−2 h−1) than the WNF method (−0.018 mg m−2 h−1).
In contrast to CO2, season-long CH4 emissions differed among GFD methods (p < 0.01). Season-long CH4 emissions were largest from the LRNF and ERNF methods (0.01 kg ha−1), which did not differ between each other nor from zero and were smallest from the LWNF (−0.51 kg ha−1) and EWNF (−0.54 kg ha−1), which did not differ between each other but were significantly less than zero (Table 2).

3.1.3. N2O Fluxes and Season-Long Emissions

Somewhat similar to CH4, but in contrast to CO2 and in contrast to that hypothesized, N2O fluxes did not differ (p > 0.05) between linear and exponential GFD models separately for the RNF and WNF datasets and did not differ (p > 0.05) between the RNF and WNF methods within the linear model, while N2O fluxes differed (p = 0.02) between the RNF and WNF methods within the exponential model (Table 1). However, N2O fluxes for all four pairs of separate comparisons (i.e., LRNF and ERNF, LWNF and EWNF, LWNF and LRNF, and EWNF and ERNF) differed over time (p < 0.01) within the growing season (Table 1).
For all four pairs of separate comparisons, N2O fluxes followed a similar general temporal pattern over the growing season (Figure 4A,B). Nitrous oxide fluxes started relatively large on DOY 129, experiencing positive and negative trends, with fluctuations cycling over a period of three weeks, and reached near-zero values in measurements at the end of the growing season on DOY 253 (Figure 4A,B). Several positive (i.e., DOY 129, 156, 192, and 212) and negative (i.e., DOY 164, 220, and 246) N2O flux peaks occurred over the course of the growing season (Figure 4A,B). The positive N2O flux peaks corresponded to furrow irrigation events that created mixed aerobic and anaerobic conditions to lead to nitrification, followed by denitrification (Figure 4A,B).
Similar to that hypothesized and similar to CH4, for the exponential model, averaged over measurement dates, the mean N2O flux was greater from RNF (0.015 mg m−2 h−1) than from the WNF method (−0.0015 mg m−2 h−1). Similar to CO2, but in contrast to CH4, season-long N2O emissions did not differ among GFD methods (p = 0.36; Table 2). Although significant variations were not present among GFD methods, season-long N2O emissions for the LRNF (0.50 kg ha−1), LWNF (0.19 kg ha−1), and ERNF (0.44 kg ha−1) GFD methods were all positive, whereas season-long N2O emissions for the EWNF (−0.09 kg ha−1) GFD method were negative (Table 2). However, no GFD model’s season-long N2O emissions differed from zero, indicating no difference in calculated season-long N2O emissions among GFD models.

3.1.4. Global Warming Potential

Similar to season-long CO2 and N2O emissions, but in contrast to season-long CH4 emissions, estimated GWP did not differ (p = 0.11) among the GFD methods evaluated (i.e., LRNF, LWNF, ERNF, and EWNF; Table 2). Global warming potential averaged 19,194 kg ha−1 across all four GFD methods (Table 2). Although no statistical difference was observed, numerically, the GWP for linear RNF and WNF GFD methods was lower than the exponential RNF and WNF GFD methods (Table 2). The lack of negative fluxes and magnitude of CO2 compared to CH4 and N2O resulted in substantially similar numerical values between RNF and WNF within linear and exponential GFD methods (Table 2).

3.2. Measurement Duration Evaluation

Two GFD models (i.e., LWNF and EWNF) were used for the statistical evaluation of the effect of measurement duration [i.e., 120, 300, 600, 900, 1200, 1500, 1800, 2100, 2400, and 2700 s, and the BAU option (120 s for CO2 and CH4, 320 s for N2O)], using a 45 s deadband duration, on GHG fluxes separately on two dates [i.e., 2 July (DOY 184) and 6 August (DOY 219), 2024] and combined across both dates.
For the LWNF GFD method, CO2 fluxes differed among measurement durations on DOY 184 (p < 0.01) and 219 (p = 0.01) but were unaffected by measurement duration when data from both sample dates were combined (Table 3). On DOY 184, which occurred early in the vegetative stage, CO2 flux was numerically largest at 300 s (5.90 mg m2 h−1) and numerically smallest at 2700 s (4.87 mg m2 h−1), while all other measurement durations, including the BAU option, were intermediate and similar to CO2 fluxes at both 300 and 2700 s (Table 4). A similar trend occurred on DOY 219, where CO2 flux was numerically largest at 120 s and for the BAU option (3.98 mg m2 h−1), which did not differ, and numerically smallest at 2700 s (3.22 mg m2 h−1), while all other measurement durations were intermediate and similar to CO2 fluxes at both 300 and 2700 s for the BAU option (Table 4).
Somewhat similar to CO2, N2O fluxes differed among measurement durations on DOY 219 (p < 0.01) and when both dates were combined (p < 0.01) but did not differ on DOY 184 (p > 0.05; Table 3). However, in contrast to CO2, on DOY 219, N2O fluxes were numerically largest at 1200, 1500, 1800, 2100, and 2400 s (0.21 mg m2 h−1), did not differ from that at 600, 900, and 2700 s, and were numerically smallest at 120 s (−1.92 mg m2 h−1) and did not differ from zero, while N2O fluxes at 300 s and for the BAU option were intermediate and similar to N2O fluxes at all other measurement durations (Table 4). Similarly, when data from both DOY 184 and 219 were combined, N2O fluxes were numerically largest at 1200, 1500, and 1800 s (0.36 mg m2 h−1), did not differ from that at 600, 900, 2100, 2400, and 2700 s, and were numerically smallest at 120 s (−1.01 mg m2 h−1) and did not differ from zero, while N2O fluxes at 300 s and for the BAU option were intermediate and similar to N2O fluxes at all other measurement durations (Table 4). Similar to that hypothesized, and in contrast to CO2 and N2O, CH4 fluxes determined with the LWNF method were unaffected by measurement duration on DOY 184 (p = 0.89), DOY 219 (p = 0.94), and when both dates were combined (p = 0.46; Table 3). In contrast to the linear model, CO2, CH4, and N2O fluxes determined with the EWNF method were unaffected (p > 0.05) by measurement duration on DOY 184 or DOY 219 or when measurement dates were combined (Table 3).

4. Discussion

4.1. Gas-Flux-Determination Methods Evaluation

4.1.1. CO2 Fluxes and Season-Long Emissions

The lack of a difference in CO2 fluxes between the WNF and RNF methods for both linear and exponential models was due to the lack of any calculated negative fluxes based on raw CO2 concentration patterns. A lack of negative CO2 fluxes was expected as the process of soil respiration is nearly always large enough to be positive, even when the soil is cold and wet [50,51]. Furthermore, the lack of negative CO2 fluxes is consistent with previous reports describing how negative CO2 fluxes are likely only the result of some sort of measurement error [52].
Additionally, the resulting fluxes for DOY 238 highlights a potential major issue with using the exponential model as the GFD method. Exponential models may be overly sensitive to influential or large concentration data points that may skew resulting gas fluxes for a particular measurement date and potentially skew an entire dataset for a longer period of time (i.e., a growing season), as occurred in the current study. A study conducted with LI-COR equipment, obtaining soil-gas measurements each second for 45 min to compare linear and non-linear models for GFD, reached a similar conclusion, stating that non-linear models were more heavily influenced by large CO2 concentrations compared to linear models [53]. Therefore, caution must be used with exponential models, and potential outliers should be thoroughly examined by observing the initial raw data and final result when processing the measured gas concentration data to determine the most appropriate gas flux to use for emissions estimates.
Although statistical differences occurred for CO2 fluxes between the linear and exponential models over time (Table 1), which supported the initial hypothesis, CO2 flux differences occurred on only 1 of 19 measurement dates in 2024 (i.e., DOY 238). Consequently, season-long CO2 emissions did not differ (p = 0.11) among GFD methods (i.e., LRNF, LWNF, ERNF, and EWNF; Table 2). Results indicate that, although significant differences occurred among GFD methods for CO2 fluxes, differences were not large nor frequent enough to significantly affect season-long CO2 emissions, despite the resulting season-long emissions from the exponential model being 1.16 times numerically larger than from the linear model (Table 2). Therefore, if solely estimating season-long CO2 emissions in cultivated, furrow-irrigated cotton on a silt–loam soil in southeast Arkansas, either the linear or exponential model appears to be appropriate.

4.1.2. CH4 Fluxes and Season-Long Emissions

The lack of significant treatment differences over time between linear and exponential models for CH4 flux calculations due to small magnitudes with large variabilities have been previously reported among upland cotton GHG studies [54,55]. Therefore, linear and exponential models applied to WNF and RNF datasets may be appropriate for determining CH4 fluxes in fine-textured upland soils.
The consistency of the CH4 flux from the RNF being greater than from the WNF model is a result of nearly all measured CH4 fluxes being recorded as negative for both the linear and exponential models. Therefore, the reassignment of negative fluxes to a small, positive number led to the difference in WNF and RNF models. Differences between RNF and WNF GFD methods indicate that negative CH4 fluxes are influential on both CH4 fluxes and season-long CH4 emissions. A significant difference between the RNF and WNF datasets indicates the possibility of errors developing and propagating over time in global CH4 emissions estimates, depending on how negative fluxes are treated during data analyses. Potential errors with CH4 flux determination and emission estimates associated with measured decreasing CH4 concentrations over time, whether real or as an artifact of the measurement procedure, agree with concerns expressed by other studies [28,56], indicating a need for further cause evaluation of negative fluxes for more accurate flux calculations and emission inventories. In the context of cotton production in upland soils, with CH4 fluxes determined using the WNF method, the presence of negative fluxes may indicate the real flux, since upland soils have often been considered CH4 sinks [57]. Therefore, the application of the WNF method for CH4 flux determinations and season-long emissions estimates may result in more accurate fluxes and emissions, provided that negative fluxes are not chamber artifacts when using the LI-COR smart chamber.

4.1.3. N2O Fluxes and Season-Long Emissions

Similar to CH4, results for N2O indicated that the choice of method to deal with negative fluxes, whether real or an artifact from the measurement technique, affected the resulting N2O flux. The significant difference in N2O fluxes between the WNF and RNF methods for the exponential model, and the absence of a difference between the WNF and RNF methods for the linear model, may result from the exponential model’s sensitivity to influential concentration data points, as previously identified for linear and exponential models for CO2. The significant difference in N2O fluxes between the RNF and WNF datasets for the exponential model and the sensitivity of the exponential model suggest that caution should be used when using the exponential model to determine N2O fluxes. Consequently, any resulting negative N2O fluxes must be analyzed critically, as soils are generally considered N2O sources [58]. Several factors, such as soil temperature, soil moisture, and fertilizer applications, have been identified to control N2O production [58,59], where the models tested may characterize the influence of environmental factors differently (Figure 4). More data are needed to fully understand the processes of N2O production and release and the causes of negative N2O fluxes as real or systematic/discrete measurement errors.
A lack of significant difference in N2O fluxes between linear and exponential models separately for the RNF and WNF datasets disagrees with former studies stating that N2O fluxes differ between linear and non-linear models [23,25,38]. However, there is a lack of studies comparing GFD methods for N2O using a closed-chamber approach with the ability to record one gas concentration measurement per second in syringe-sampled approaches with fewer gas concentration measurements. Therefore, using either a linear or exponential model for N2O flux determination may be appropriate for the LI-COR smart chamber in upland soils, with consideration of potential differences arising from the EWNF and ERNF methods. The lack of significant differences between GFD models for season-long N2O emissions showed that each of the four GFD models evaluated (i.e., LRNF, LWNF, ERNF, and EWNF) may be appropriate for use when estimating season-long N2O emissions from furrow-irrigated cotton production in an upland, fine-textured soil in southeast Arkansas.

4.1.4. Global Warming Potential

The lack of significant difference in GWP between GFD methods indicates that each GFD method (i.e., LRNF, LWNF, ERNF, and EWNF) is appropriate for estimating GWP in upland cotton production. To the author’s knowledge, no field studies have been conducted comparing GFD methods and GWP estimations, particularly with regard to fitting gas concentrations over time with linear and/or exponential models, with or without the inclusion of negative fluxes. Further investigation of potential GFD-method effects on GHG fluxes and season-long emissions and GWP estimations in other cropping systems with varying GHG flux magnitudes and management practices should be conducted.

4.2. Measurement Duration Evaluation

The gradual numeric decrease in CO2 flux as measurement duration increased agreed with the laws of gas diffusion [39], where a gas concentration will generally increase in a chamber until a maximum concentration is achieved, then tend to reserve the gradient and diffuse back into the soil, decreasing the headspace gas concentration. The Soil Flux Pro software from LI-COR does not account for the diffusion-gradient change in a chamber’s headspace during flux determinations and therefore recommends maintaining a relatively short measurement duration for CO2 measurements [18]. Based on the separate measurement date evaluations during different growth stages, but in contrast to that hypothesized, field data from cotton support the recommendation to keep CO2 concentration measurement between 120 to 300 s when using a linear GFD model to obtain the maximum CO2 flux before headspace CO2 concentrations start to decrease by diffusion outside of the chamber, which agreed with LI-COR’s 120-s CO2 measurement duration recommendation [41].
The results indicate that 300 to 2700 s measurement durations and the BAU option are appropriate for N2O flux determinations with the LWNF option for upland soils. A 300 to 2700 s measurement duration for N2O flux determinations with the LWNF method agrees with previous studies, where a measurement duration is often greater than 120 s [40,60]. However, statistically similar N2O fluxes for the LWNF method varied from negative to positive magnitudes, thus highlighting the potential discrepancies in flux determinations and the present difficulties in understanding soil processes leading to N2O production and release. Consequently, a need for further investigation of soil processes affecting N2O production and release is still required to determine the optimal N2O-measurement duration.
The lack of CH4 flux differences among measurement durations likely results from the small CH4-flux magnitudes throughout the growing season. Consequently, a 120 to 2700 s measurement duration appears appropriate for CH4 flux determinations using the LWNF method in upland soils when CH4 fluxes are, or are expected to be, low. However, LI-COR recommends CH4 concentration data collection at a 90 to 180 s measurement duration to avoid gas build-up within the chamber’s headspace to potentially affect the gradient [41].
The absence of GHG-flux differences among measurement durations using the EWNF method may indicate that the exponential model is heavily influenced by concentrations measured at short measurement durations and, therefore, does not substantially deviate from longer measurement durations. Assuming the EWNF method is only influenced by GHG concentrations measured during short measurement durations, short measurement durations (i.e., <300 s) appear sufficient when applying the EWNF method to determine CO2, CH4, and N2O fluxes in upland soils.
In a study that measured CO2 fluxes with a closed-chamber technique on sandy clay loam, sandy loam, and peat soils using a 45 min measurement duration with one concentration measurement per second, concluded that CO2 fluxes determined with the linear model were an underestimate compared to CO2 fluxes determined with the exponential model [53].

4.3. Implications

The current absence of systematic evaluations of GFD methods, measurement, and deadband duration, particularly with LI-COR equipment, indicates a need for the evaluation of GHG measurement methodologies. Without consistent and reliable methods for obtaining GHG data, differences in calculated GHG fluxes, season-long emissions, and GWP likely occur across studies, leading to unreliable data comparisons and inaccurate GHG emissions inventories. As evident by the current study, significant differences in the GFD method with respect to GHG fluxes, season-long emissions, and GWP occurred, leading to the conclusion that the use of either a linear or exponential model will affect the resulting GHG flux when influential data points are obtained, and the inclusion of negative flux values impact season-long emissions, particularly for CH4 in upland soils. The evaluation of measurement duration also provides ranges of appropriate measurement and deadband sequences for consistent measurement methodology.

5. Conclusions

This study generated high-quality datasets from simultaneous, in-field measurements of CO2, CH4, and N2O over time using state-of-the-art, field-portable analyzers to formally evaluate data-analysis options. Specifically, the effects of the GFD method (i.e., LRNF, LWNF, ERNF, and EWNF) on CO2, CH4, and N2O fluxes and season-long emissions and GWP were evaluated over the 2024 growing season in a furrow-irrigated cotton production system in southeast Arkansas using LI-COR equipment (i.e., LI-7810, LI-7820, and LI-8200-01S, LI-COR, Lincoln, NE, USA).
In partial accordance with the hypothesis, the exponential-calculated CO2 flux was greater than the linear flux, but for only one data point. However, there was no difference in season-long CO2 emissions between GFD methods, concluding that both linear and exponential models for CO2 flux evaluation may be appropriate. However, caution must be used, as exponential models may be sensitive to abnormal gas concentration data points and may not represent the true flux value. In contrast, no differences between linear and exponential models occurred for CH4 and N2O fluxes, indicating that both linear and exponential models are appropriate, with the application of proper caution.
Methane fluxes differed between RNF and WNF for both linear and exponential GFD methods, where the RNF was consistently greater than the WNF, which was similar to what was hypothesized. In addition, season-long CH4 emissions from the RNF and WNF GFD methods differed due to the predominance of negative fluxes measured throughout the growing season, indicating that a discrepancy in emission inventories may occur if only using the RNF method.
In contrast to what was hypothesized, N2O fluxes did not differ between linear and exponential GFD methods and only differed between ERNF and EWNF, where N2O fluxes from the ERNF methods were consistently greater than from the EWNF method. The difference for only exponential fluxes with RNF and WNF demonstrated that the exponential model is sensitive to influential gas concentration data points and may vary more than fluxes from the linear model. No differences in season-long N2O emissions occurred for any GFD model, further indicating that linear and exponential and WNF or RNF combinations may be appropriate for N2O GFD methods, considering appropriate caution with the exponential model. However, for the consistency and reliability of flux evaluation, a linear model is recommended for use.
Partially supporting what was hypothesized, it was concluded CO2, CH4, and N2O measurement durations should be kept short (i.e., <300 s) to avoid the build-up of gases in the chamber’s headspace when obtaining one GHG measurement per second. It was also concluded that linear fluxes tend to be greater than exponential fluxes, particularly for CO2, which was contrary to what was hypothesized.
As GHG fluxes and emissions vary by crop, soil texture, and land management practices, the appropriate GFD method and measurement duration may also vary, indicating a need for further research on various agroecosystems over multiple years. Different GHG measurement instruments may also result in discrepancies with appropriate measurement durations for obtaining GHG data; therefore, further studies should also be conducted on instruments other than from LI-COR in an attempt to work toward a standardized GFD method. Furthermore, future studies should consider investigating error propagation for flux calculations across the different GFD methods, along with trends in the coefficient of variation of the slope for the different methods in order to improve standard operating procedures for GHG analyses.

Author Contributions

Conceptualization, K.B., D.D.L. and C.S.; Methodology, K.B., D.D.L., C.S. and J.B.; Formal Analysis, K.B., D.D.L., C.S. and J.B.; Investigation, K.B., D.D.L., C.S. and J.B.; Resources, K.B. and M.D.; Data Curation, K.B., D.D.L., C.S. and J.B.; Writing—Original Draft Preparation, C.S.; Writing—Review and Editing, K.B., D.D.L., J.B., M.D., L.W. and K.G.; Supervision, K.B., D.D.L. and M.D. Project Administration, K.B. and M.D.; Funding Acquisition, K.B. and M.D. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by USDA-NIFA Farm Production and Conservation program, award number NR233A750004G043.

Data Availability Statement

The data presented in this study are available upon request from the corresponding author.

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:
ANOVAAnalysis of Variance
BAUBusiness-as-usual
CH4Methane
CO2Carbon Dioxide
CCCover Crop
DOYDay of Year
ECElectrical Conductivity
EExponential
GFDGas-flux-determination
GHGGreenhouse Gas
GWPGlobal Warming Potential
IGRAInfrared Gas Analyzer
IPCCIntergovernmental Panel on Climate Change
LLinear
LMRVLower Mississippi River Valley
N2ONitrous Oxide
PVCPolyvinyl Chloride
RNFReplacing Negative Fluxes
WNFWith Negative Fluxes

References

  1. Canadell, J.G.; Monteiro, P.M.S.; Costa, M.H.; Cotrim da Cunha, L.; Cox, P.M.; Eliseev, A.V.; Henson, S.; Ishii, M.; Jaccard, S.; Koven, C.; et al. Global carbon and other biogeochemical cycles and feedbacks. In Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change; Masson-Delmotte, V., Zhai, P., Pirani, A., Connors, S.L., Péan, C., Berger, S., Caud, N., Chen, Y., Goldfarb, L., Gomis, M.I., et al., Eds.; Cambridge University Press: Cambridge, UK; New York, NY, USA, 2021; pp. 673–816. [Google Scholar]
  2. Jia, G.; Shevliakova, E.; Artaxo, P.; De Noblet-Ducoudré, N.; Houghton, R.; House, J.; Kitajima, K.; Lennard, C.; Popp, A.; Sirin, A.; et al. Land–climate interactions. In Climate Change and Land: An IPCC Special Report on Climate Change, Desertification, Land Degradation, Sustainable Land Management, Food Security, and Greenhouse Gas Fluxes in Terrestrial Ecosystems; Shukla, P.R., Skea, J., Calvo Buendia, E., Masson-Delmotte, V., Pörtner, H.-O., Roberts, D.C., Zhai, P., Slade, R., Connors, S., van Diemen, R., et al., Eds.; Cambridge University Press: Cambridge, UK, 2019; pp. 131–248. [Google Scholar]
  3. Matthias, A.D.; Blackmer, A.M.; Bremner, J.M. A simple chamber technique for field measurement of emissions of nitrous oxide from soils. J. Environ. Qual. 1980, 9, 251–256. [Google Scholar] [CrossRef]
  4. Parkinson, K.J. An improved method for measuring soil respiration in the field. J. Appl. Ecol. 1981, 18, 221–228. [Google Scholar] [CrossRef]
  5. Hutchinson, G.L.; Mosier, A.R. Improved soil cover method for field measurement of nitrous oxide fluxes. Soil Sci. Soc. Am. J. 1981, 45, 311–316. [Google Scholar] [CrossRef]
  6. Crill, P.M. Seasonal patterns of methane uptake and carbon dioxide release by a temperate woodland soil. Glob. Biogeochem. Cycles 1991, 5, 319–334. [Google Scholar] [CrossRef]
  7. Scott, A.; Crichton, I.; Ball, B.C. Long-term monitoring of soil gas fluxes with closed chambers using automated and manual systems. J. Environ. Qual. 1999, 28, 1637–1643. [Google Scholar] [CrossRef]
  8. Rochette, P.; Hutchinson, G.L. Measurement of soil respiration in situ: Chamber techniques. In Agronomy Monographs; Hatfield, J.L., Baker, J.M., Eds.; USDA-ARS/UNL: Lincoln, NE, USA, 2005; pp. 247–286. [Google Scholar]
  9. Rochette, P.; Ellert, B.; Gregorich, E.G.; Desjardins, R.L.; Pattey, E.; Lessard, R.; Johnson, B.G. Description of a dynamic closed chamber for measuring soil respiration and its comparison with other techniques. Can. J. Soil Sci. 1997, 77, 195–203. [Google Scholar] [CrossRef]
  10. Gao, F.; Yates, S.R.; Yates, M.V.; Gan, J.; Ernst, F.F. Design, fabrication, and application of a dynamic chamber for measuring gas emissions from soil. Environ. Sci. Technol. 1997, 31, 148–153. [Google Scholar] [CrossRef]
  11. Reichman, R.; Rolston, D.E. Design and performance of a dynamic gas flux chamber. J. Environ. Qual. 2002, 31, 1774–1781. [Google Scholar] [CrossRef]
  12. LI-COR Environmental. LI-7810 CH4/CO2/H2O Trace Gas Analyzer Instruction Manual. Available online: https://licor.app.boxenterprise.net/s/zsce2z88fgkfpbiguqi0k3g537jllsl7 (accessed on 15 July 2024).
  13. LI-COR Environmental. LI-7820 N2O/H2O Trace Gas Analyzer Instruction Manual. Available online: https://licor.app.boxenterprise.net/s/xg9i6bs12w8h8y3g1wqxltuxlzr4m6qz (accessed on 15 July 2024).
  14. Bohra, R. Greenhouse Gas Fluxes Across Climatic Gradients in Mountain National Parks of the Iberian Peninsula. Master’s Thesis, University of Eastern Finland, Kuopio, Finland, 2023. [Google Scholar]
  15. Hopple, A.M.; Pennington, S.C.; Megonigal, J.P.; Bailey, V.; Bond-Lamberty, B. Root and microbial soil CO2 and CH4 fluxes respond differently to seasonal and episodic environmental changes in a temperate forest. J. Geophys. Res.-Biogeosci. 2023, 128, e2022JG007233. [Google Scholar] [CrossRef]
  16. Ingle, R.; Habib, W.; Connolly, J.; McCorry, M.; Barry, S.; Saunders, M. Upscaling methane fluxes from peatlands across a drainage gradient in Ireland using PlanetScope imagery and machine learning tools. Sci. Rep. 2023, 13, 11997. [Google Scholar] [CrossRef]
  17. Koschorreck, M.; Kamjunke, N.; Koedel, U.; Rode, M.; Schuetze, C.; Bussmann, I. Diurnal versus spatial variability of greenhouse gas emissions from an anthropogenically modified lowland river in Germany. Biogeosciences 2024, 21, 1613–1628. [Google Scholar] [CrossRef]
  18. Hupp, J. Theory and Measurement of Trace Gas Fluxes from Soil; Training Manual as a Power Point Presentation; LI-COR Environmental, Inc.: Lincoln, NE, USA, 2022. [Google Scholar]
  19. Tenny, K.M.; Cooper, J.S. Ideal Gas Behavior; StatsPearl Publishing: Treasure Island, FL, USA, 2025. [Google Scholar]
  20. Livingston, G.P.; Hutchinson, G.L.; Spartalian, K. Diffusion theory improves chamber-based measurements of trace gas emissions. Geophys. Res. Lett. 2005, 32, L24817. [Google Scholar] [CrossRef]
  21. Livingston, G.P.; Hutchinson, G.L.; Spartalian, K. Trace gas emission in chambers: A non-steady-state diffusion model. Soil Sci. Soc. Am. J. 2006, 70, 1459–1469. [Google Scholar] [CrossRef]
  22. Kutzbach, L.; Schneider, J.; Sachs, T.; Giebels, M.; Nykänen, H.; Shurpali, N.J.; Martikainen, P.J.; Alm, J.; Wilmking, M. CO2 flux determination by closed-chamber methods can be seriously biased by inappropriate application of linear regression. Biogeosciences 2007, 4, 1005–1025. [Google Scholar] [CrossRef]
  23. Venterea, R.T.; Spokas, K.A.; Baker, J.M. Accuracy and precision analysis of chamber-based nitrous oxide gas flux estimates. Soil Sci. Soc. Am. J. 2009, 73, 1087–1093. [Google Scholar] [CrossRef]
  24. Venterea, R.T.; Petersen, S.O.; de Klein, C.A.M.; Pedersen, A.R.; Noble, A.D.L.; Rees, R.M.; Gamble, J.D.; Parkin, T.B. Global Research Alliance N2O chamber methodology guidelines: Flux calculations. J. Environ. Qual. 2020, 49, 1141–1155. [Google Scholar] [CrossRef]
  25. Silva, J.P.; Lasso, A.; Lubberding, H.J.; Pena, M.R.; Gijzen, H.J. Biases in greenhouse gases static chambers measurements in stabilization ponds: Comparison of flux estimation using linear and non-linear models. Atmos. Environ. 2015, 109, 130–138. [Google Scholar] [CrossRef]
  26. Richardson, C.J.; Flanagan, N.E.; Wang, H.; Ho, M. Annual carbon sequestration and loss rates under altered hydrology and fire regimes in southeastern USA pocosin peatlands. Glob. Change Biol. 2022, 28, 6370–6384. [Google Scholar] [CrossRef]
  27. Ruis, S.J.; Blanco-Canqui, H.; Jasa, P.J.; Jin, V.L. No-till farming and greenhouse gas fluxes: Insights from literature and experimental data. Soil Tillage Res. 2022, 220, 105359. [Google Scholar] [CrossRef]
  28. Courtois, E.A.; Stahl, C.; Burban, B.; Van den Berge, J.; Berveiller, D.; Bréchet, L.; Soong, J.L.; Arriga, N.; Peñuelas, J.; Janssens, I.A. Automatic high-frequency measurements of full soil greenhouse gas fluxes in a tropical forest. Biogeosciences 2019, 16, 785–796. [Google Scholar] [CrossRef]
  29. Parkin, T.B.; Venterea, R.T. Chamber-Based Trace Gas Flux Measurements. In Sampling Protocols; Follett, R.F., Ed.; USDA-ARS: Washington, DC, USA, 2010; pp. 3.1–3.39. [Google Scholar]
  30. Rogers, C.W.; Brye, K.R.; Norman, R.J.; Gbur, E.E.; Mattice, J.D.; Parkin, T.B.; Roberts, T.L. Methane emissions from drill-seeded, delayed-flood rice production on a silt-loam soil in Arkansas. J. Environ. Qual. 2013, 42, 1059–1069. [Google Scholar] [CrossRef]
  31. Rogers, C.W.; Brye, K.R.; Smartt, A.D.; Norman, R.J.; Gbur, E.E.; Evans-White, M.A. Cultivar and previous crop effects on methane emissions from drill-seeded, delayed-flood rice production on a silt-loam soil. Soil Sci. 2014, 179, 28–36. [Google Scholar] [CrossRef]
  32. Rogers, C.W.; Smartt, A.D.; Brye, K.R.; Norman, R.J. Nitrogen source effects on methane emissions from drill-seeded, delayed-flood rice production. Soil Sci. 2017, 182, 9–17. [Google Scholar] [CrossRef]
  33. Smartt, A.D.; Brye, K.R.; Rogers, C.W.; Norman, R.J.; Gbur, E.E.; Hardke, J.T.; Roberts, T.L. Previous crop and cultivar effects on methane emissions from drill-seeded, delayed-flood rice grown on a clay soil. Appl. Environ. Soil Sci. 2016, 2016, 9542361. [Google Scholar] [CrossRef]
  34. Slayden, J.M.; Brye, K.R.; Della Lunga, D.; Henry, C.G.; Wood, L.S.; Lessner, D.J. Site position and tillage treatment effects on nitrous oxide emissions from furrow-irrigated rice on a silt-loam Alfisol in the Mid-south, USA. Geoderma Reg. 2022, 28, e00491. [Google Scholar] [CrossRef]
  35. United States Department of Agriculture, Foreign Agriculture Service. Cotton Explorer. Available online: https://ipad.fas.usda.gov/cropexplorer/cropview/commodityView.aspx?cropid=2631000&sel_year=2023&rankby=Production (accessed on 12 August 2024).
  36. MacDonald, S.; Lanclos, K.; Meyer, L.; Soley, G. The world and United States cotton outlook. In Proceedings of the U.S. Department of Agriculture 100th Annual Agricultural Outlook Forum, Arlington, VA, USA, 15–16 February 2024. [Google Scholar]
  37. United States Department of Agriculture, Economic Research Service. Cotton Sector at a Glance. Available online: https://www.ers.usda.gov/topics/crops/cotton-and-wool/cotton-sector-at-a-glance/ (accessed on 13 August 2024).
  38. Anthony, W.H.; Hutchinson, G.L.; Livingston, G.P. Chamber measurement of soil-atmosphere gas exchange: Linear vs. diffusion-based flux models. Soil Sci. Soc. Am. J. 1995, 59, 1308–1310. [Google Scholar] [CrossRef]
  39. Healy, R.W.; Striegl, R.G.; Russell, T.F.; Hutchinson, G.L.; Livingston, G.P. Numerical evaluation of static-chamber measurements of soil-atmosphere gas exchange: Identification of physical processes. Soil Sci. Soc. Am. J. 1996, 60, 740–747. [Google Scholar] [CrossRef]
  40. LI-COR Environmental. High Performance CO2, CO2 Isotope, NH3, CH4, and N2O Gas Analyzers for the Field, the Lab, and Mobile Applications. Available online: https://www.compact-industrial.ro/wp-content/uploads/2024/02/78xx.pdf (accessed on 15 August 2024).
  41. LI-COR Environmental. Support: SoilFluxPro 5 Software. Available online: https://www.licor.com/env/support/SoilFluxPro/topics/file-management-bar.html#Mappingdata (accessed on 9 June 2025).
  42. United States Department of Agriculture, Natural Resource Conservation Service. Land Resource Regions and Major Land Resource Areas of the United States, the Caribbean, and the Pacific Basin. Available online: https://www.nrcs.usda.gov/sites/default/files/2022-10/AgHandbook296_text_low-res.pdf (accessed on 4 September 2024).
  43. United States Department of Agriculture, National Cooperative Soil Survey. Hebert Series. Available online: https://soilseries.sc.egov.usda.gov/OSD_Docs/H/HEBERT.html (accessed on 4 September 2024).
  44. United States Department of Agriculture, Natural Resource Conservation Service. Web Soil Survey. Available online: https://websoilsurvey.nrcs.usda.gov/app/ (accessed on 4 September 2024).
  45. National Oceanic & Atmospheric Administration, National Centers for Environmental Information. U.S. Climate Normals Quick Access. Available online: https://www.ncei.noaa.gov/access/us-climate-normals/#dataset=normals-monthly&timeframe=30&location=AR&station=USC00032148 (accessed on 4 September 2024).
  46. Forster, P.; Storelvmo, T.; Armour, K.; Collins, W.; Dufresne, J.-L.; Frame, D.; Lunt, D.J.; Mauritsen, T.; Palmer, M.D.; Watanabe, M.; et al. The Earth’s energy budget, climate feedbacks, and climate sensitivity. In Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change; Masson-Delmotte, V., Zhai, P., Pirani, A., Connors, S.L., Péan, C., Berger, S., Caud, N., Chen, Y., Goldfarb, L., Gomis, M.I., et al., Eds.; Cambridge University Press: Cambridge, UK; New York, NY, USA, 2021; pp. 923–1054. [Google Scholar]
  47. Firth, A.G.; Brooks, J.P.; Locke, M.A.; Morin, D.J.; Brown, A.; Baker, B.H. Dynamics of soil organic carbon and CO2 flux under cover crop and no-till management in soybean cropping systems of the mid-south (USA). Environments 2022, 9, 109. [Google Scholar] [CrossRef]
  48. Hu, J.; Lei, F.; Miles, D.M.; Adeli, A.; Brooks, J.P.; Smith, R.; Podrebarac, F.A.; Li, X.; Moorhead, R. Effects of cover crops and soil amendments on soil CO2 flux in a Mississippi corn cropping system on upland soil. Environments 2023, 10, 19. [Google Scholar] [CrossRef]
  49. Ma, L.; Kong, F.; Lv, X.; Wang, Z.; Zhou, Z.; Meng, Y. Responses of greenhouse gas emissions to different straw management methods with the same amount of carbon input in cotton field. Soil Tillage Res. 2021, 213, 105126. [Google Scholar] [CrossRef]
  50. Yuste, C.J.; Baldocchi, D.D.; Gershenson, A.; Goldstein, A.; Misson, L.; Wong, S. Microbial soil respiration and its dependency on carbon inputs, soil temperature and moisture. Glob. Change Biol. 2007, 13, 2018–2035. [Google Scholar] [CrossRef]
  51. Risk, D.; Kellman, L.; Beltrami, H. Carbon dioxide in soil profiles: Production and temperature dependence. Geophys. Res. Lett. 2002, 29, 1087. [Google Scholar] [CrossRef]
  52. Collier, S.M.; Ruark, M.D.; Oates, L.G.; Jokela, W.E.; Dell, C.J. Measurement of greenhouse gas flux from agricultural soils using static chambers. J. Vis. Exp. 2014, 90, 52210. [Google Scholar]
  53. Kandel, T.P.; Lærke, P.E.; Elsgaard, L. Effect of chamber enclosure time on soil respiration flux: A comparison of linear and non-linear flux calculation methods. Atmos. Environ. 2016, 141, 245–254. [Google Scholar] [CrossRef]
  54. Tian, Z.; Wang, J.J.; Liu, S.; Zhang, Z.; Dodla, S.K.; Myers, G. Application effects of coated urea and urease and nitrification inhibitors on ammonia and greenhouse gas emissions from a subtropical cotton field of the Mississippi delta region. Sci. Total Environ. 2015, 533, 329–338. [Google Scholar] [CrossRef]
  55. Watts, D.B.; Runion, G.B.; Smith Nannenga, K.W.; Torbert, H.A. Impacts of enhanced-efficiency nitrogen fertilizers on greenhouse gas emissions in a coastal plain soil under cotton. J. Environ. Qual. 2015, 44, 1699–1710. [Google Scholar] [CrossRef]
  56. Cueva, A.; Volkmann, H.M.; van Haren, J.; Troch, P.A.; Meredith, L.K. Reconciling negative soil CO2 fluxes: Insights from a large-scale experimental hillslope. Soil Syst. 2019, 3, 10. [Google Scholar] [CrossRef]
  57. Le Mer, J.; Roger, P. Production, oxidation, emission and consumption of methane by soils: A review. Eur. J. Soil Biol. 2001, 37, 25–50. [Google Scholar] [CrossRef]
  58. Chapuis-Lardy, L.; Wrage, N.; Metay, A.; Chotte, J.-L.; Bernoux, M. Soils, a sink for N2O? A review. Glob. Change Biol. 2007, 13, 1–17. [Google Scholar] [CrossRef]
  59. van Groenigen, J.W.; Huygens, D.; Boeckx, P.; Kuyper, T.W.; Lubbers, I.M.; Rütting, T.; Groffman, P.M. The soil N cycle: New insights and key challenges. Soil 2015, 1, 235–256. [Google Scholar] [CrossRef]
  60. Parker, D.; Casey, K.; Todd, R.; Waldrip, H.; Marek, G.; Auvermann, B.; Marek, T.; Webb, K.; Willis, W.; Pemberton, B.; et al. Improved chamber systems for rapid, real-time nitrous oxide emissions from manure and soil. Trans. ASABE 2017, 60, 1235–1258. [Google Scholar] [CrossRef]
Agriengineering 07 00213 g001
Figure 1. In-field locations of greenhouse gas sampling base collars in a furrow-irrigated cotton field in Desha County, Arkansas. The red, rectangular border encompasses the 5.6 m wide and 49 m long study area that is 46 m downslope from the north edge of the field. Green circles represent the specific base collar locations that are approximately 2.8 m between adjacent base collars.
Figure 1. In-field locations of greenhouse gas sampling base collars in a furrow-irrigated cotton field in Desha County, Arkansas. The red, rectangular border encompasses the 5.6 m wide and 49 m long study area that is 46 m downslope from the north edge of the field. Green circles represent the specific base collar locations that are approximately 2.8 m between adjacent base collars.
Agriengineering 07 00213 g001
Agriengineering 07 00213 g002
Figure 2. Carbon dioxide (CO2) fluxes for linear with reassigned negative fluxes (LRNF) and exponential with reassigned negative fluxes (ERNF) over time from furrow-irrigated cotton grown in 2024 on a silt–loam soil in southeast Arkansas. All CO2 fluxes were significantly greater (p < 0.05) than a flux of zero. The standard error from Tukey’s mean separation analysis was 0.14 mg m−2 h−1.
Figure 2. Carbon dioxide (CO2) fluxes for linear with reassigned negative fluxes (LRNF) and exponential with reassigned negative fluxes (ERNF) over time from furrow-irrigated cotton grown in 2024 on a silt–loam soil in southeast Arkansas. All CO2 fluxes were significantly greater (p < 0.05) than a flux of zero. The standard error from Tukey’s mean separation analysis was 0.14 mg m−2 h−1.
Agriengineering 07 00213 g002
Agriengineering 07 00213 g003
Figure 3. Methane (CH4) fluxes over time, averaged across the linear with reassigned negative fluxes (LRNF) and exponential with reassigned negative fluxes (ERNF) gas-flux-determination methods, from furrow-irrigated cotton grown in 2024 on a silt–loam soil in southeast Arkansas. All CH4 fluxes from the LRNF and ERNF methods differed (p < 0.05) from a flux of zero. The standard error from Tukey’s mean separation analysis was 0.35 mg m−2 h−1.
Figure 3. Methane (CH4) fluxes over time, averaged across the linear with reassigned negative fluxes (LRNF) and exponential with reassigned negative fluxes (ERNF) gas-flux-determination methods, from furrow-irrigated cotton grown in 2024 on a silt–loam soil in southeast Arkansas. All CH4 fluxes from the LRNF and ERNF methods differed (p < 0.05) from a flux of zero. The standard error from Tukey’s mean separation analysis was 0.35 mg m−2 h−1.
Agriengineering 07 00213 g003
Agriengineering 07 00213 g004
Figure 4. Nitrous oxide (N2O) over time, averaged across pairs of gas-flux-determination methods [i.e., linear and exponential with negative fluxes (LWNF and EWNF), linear and exponential with reassigned negative fluxes (LRNF and ERNF), LWNF and LRNF, and EWNF and ERNF] from furrow-irrigated cotton grown in 2024 on a silt–loam soil in southeast Arkansas. The N2O fluxes for LRNF and ERNF all differed (p < 0.05) from a flux of zero. Nitrous oxide fluxes for ERNF and EWNF from −0.025 to 0.034 mg m−2 h−1 did not differ (p > 0.05) from a flux of zero. Nitrous oxide fluxes for EWNF and LWNF from −0.027 to 0.035 mg m−2 h−1 did not differ (p > 0.05) from a flux of zero. All N2O fluxes for LRNF and LWNF greater than 0.020 mg m−2 h−1 differed (p < 0.05) from a flux of zero. The standard error from Tukey’s mean separation analysis was 0.14 mg m−2 h−1 for Panel (A) and 0.01 mg m−2 h−1 for Panel (B).
Figure 4. Nitrous oxide (N2O) over time, averaged across pairs of gas-flux-determination methods [i.e., linear and exponential with negative fluxes (LWNF and EWNF), linear and exponential with reassigned negative fluxes (LRNF and ERNF), LWNF and LRNF, and EWNF and ERNF] from furrow-irrigated cotton grown in 2024 on a silt–loam soil in southeast Arkansas. The N2O fluxes for LRNF and ERNF all differed (p < 0.05) from a flux of zero. Nitrous oxide fluxes for ERNF and EWNF from −0.025 to 0.034 mg m−2 h−1 did not differ (p > 0.05) from a flux of zero. Nitrous oxide fluxes for EWNF and LWNF from −0.027 to 0.035 mg m−2 h−1 did not differ (p > 0.05) from a flux of zero. All N2O fluxes for LRNF and LWNF greater than 0.020 mg m−2 h−1 differed (p < 0.05) from a flux of zero. The standard error from Tukey’s mean separation analysis was 0.14 mg m−2 h−1 for Panel (A) and 0.01 mg m−2 h−1 for Panel (B).
Agriengineering 07 00213 g004
Table 1. Analysis of variance summary of the effects of time (i.e., measurement week), treatment [i.e., gas-flux-determination (GFD) method], and their interaction for comparisons between pairs of GFD methods [i.e., linear with reassigned negative fluxes (LRNF), linear with negative fluxes (LWNF), exponential with reassigned negative fluxes (ERNF) and exponential with negative fluxes (EWNF)], separated by gas [i.e., carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O)], from furrow-irrigated cotton grown in 2024 on a silt–loam soil in southeast Arkansas.
Table 1. Analysis of variance summary of the effects of time (i.e., measurement week), treatment [i.e., gas-flux-determination (GFD) method], and their interaction for comparisons between pairs of GFD methods [i.e., linear with reassigned negative fluxes (LRNF), linear with negative fluxes (LWNF), exponential with reassigned negative fluxes (ERNF) and exponential with negative fluxes (EWNF)], separated by gas [i.e., carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O)], from furrow-irrigated cotton grown in 2024 on a silt–loam soil in southeast Arkansas.
GFD Method ComparisonSource of VariationCO2CH4N2O
p
LRNF and ERNFTime<0.01<0.01<0.01
Treatment0.020.990.73
    Time x treatment0.040.990.99
LWNF and EWNFTime<0.010.07<0.01
Treatment0.010.720.27
    Time x treatment0.050.990.99
LWNF and LRNFTime<0.010.89<0.01
Treatment0.99<0.010.13
    Time x treatment0.990.520.99
EWNF and ERNFTime<0.010.86<0.01
Treatment0.99<0.010.02
    Time x treatment0.990.510.88
Table 2. Analysis of variance summary of the effect of the gas-flux-determination (GFD) method [i.e., linear with reassigned negative fluxes (LRNF), linear with negative fluxes (LWNF), exponential with reassigned negative fluxes (ERNF), and exponential with negative fluxes (EWNF)] on season-long gas [i.e., carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O)] emissions and season-long global warming potential (GWP) from furrow-irrigated cotton grown in 2024 on a silt–loam soil in southeast Arkansas.
Table 2. Analysis of variance summary of the effect of the gas-flux-determination (GFD) method [i.e., linear with reassigned negative fluxes (LRNF), linear with negative fluxes (LWNF), exponential with reassigned negative fluxes (ERNF), and exponential with negative fluxes (EWNF)] on season-long gas [i.e., carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O)] emissions and season-long global warming potential (GWP) from furrow-irrigated cotton grown in 2024 on a silt–loam soil in southeast Arkansas.
GFD MethodCO2 (kg ha−1)CH4 (kg ha−1)N2O (kg ha−1)GWP (kg ha−1)
LRNF17,7040.01 a†0.5017838
LWNF17,704−0.51 b0.1917741
ERNF20,5610.01 a0.4420678
EWNF20,561−0.54 b−0.0920522
p-value0.11<0.010.360.11
† Means within a column with different lower-case letters are different at p < 0.05.
Table 3. Analysis of variance summary of the effect of measurement duration [i.e., 120, 300, 600, 900, 1200, 1500, 1800, 2100, 2400, and 2700 s, and the business-as-usual (BAU) option (120 s for carbon dioxide (CO2) and methane (CH4) and 320 s for nitrous oxide (N2O))], using a 45 s deadband, separated by gas-flux-determination (GFD) method [i.e., linear with negative fluxes (LWNF) and exponential with negative fluxes (EWNF)], by gas [i.e., carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O)], and by two different measurement dates by day of year (DOY) and data combined across the two measurement dates, from furrow-irrigated cotton grown in 2024 on a silt–loam soil in southeast Arkansas.
Table 3. Analysis of variance summary of the effect of measurement duration [i.e., 120, 300, 600, 900, 1200, 1500, 1800, 2100, 2400, and 2700 s, and the business-as-usual (BAU) option (120 s for carbon dioxide (CO2) and methane (CH4) and 320 s for nitrous oxide (N2O))], using a 45 s deadband, separated by gas-flux-determination (GFD) method [i.e., linear with negative fluxes (LWNF) and exponential with negative fluxes (EWNF)], by gas [i.e., carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O)], and by two different measurement dates by day of year (DOY) and data combined across the two measurement dates, from furrow-irrigated cotton grown in 2024 on a silt–loam soil in southeast Arkansas.
GFD
Method		Date 1 (DOY 184)Date 2 (DOY 219)Dates Combined
CO2CH4N2OCO2CH4N2OCO2CH4N2O
p
LWNF<0.010.890.990.010.94<0.010.780.46<0.01
EWNF0.960.480.980.160.940.170.980.270.21
Table 4. Effect of measurement duration [i.e., 120, 300, 600, 900, 1200, 1500, 1800, 2100, 2400, and 2700 s, and the business-as-usual (BAU) option (i.e., 120 s for carbon dioxide (CO2) and methane (CH4) and 320 s for nitrous oxide (N2O))], using a 45 s deadband, separated by gas (i.e., CO2 and N2O) for the linear with negative flux (LWNF) gas-flux-determination method and separated by two measurement dates by day of year (DOY) and data combined across measurement dates, from furrow-irrigated cotton grown in 2024 on a silt–loam soil in southeast Arkansas.
Table 4. Effect of measurement duration [i.e., 120, 300, 600, 900, 1200, 1500, 1800, 2100, 2400, and 2700 s, and the business-as-usual (BAU) option (i.e., 120 s for carbon dioxide (CO2) and methane (CH4) and 320 s for nitrous oxide (N2O))], using a 45 s deadband, separated by gas (i.e., CO2 and N2O) for the linear with negative flux (LWNF) gas-flux-determination method and separated by two measurement dates by day of year (DOY) and data combined across measurement dates, from furrow-irrigated cotton grown in 2024 on a silt–loam soil in southeast Arkansas.
Measurement
Duration		Date 1 (DOY 184)Date 2 (DOY 219)Date 2 (DOY 219)Dates Combined
CO2
(mg m−2 h−1)		CO2
(mg m−2 h−1)		N2O
(mg m−2 h−1)		N2O
(mg m−2 h−1)
BAU5.81 ab†3.98 a−0.58 ab−0.15 ab
120 s5.81 ab3.98 a−1.92 b−1.01 b
300 s5.90 a3.85 ab−0.62 ab−0.17 ab
600 s5.69 ab3.74 ab0.08 a0.26 a
900 s5.52 ab3.64 ab0.17 a0.34 a
1200 s5.39 ab3.55 ab0.21 a0.36 a
1500 s5.28 ab3.47 ab0.21 a0.36 a
1800 s5.17 ab3.40 ab0.21 a0.36 a
2100 s5.06 ab3.34 ab0.21 a0.35 a
2400 s4.97 ab3.28 ab0.21 a0.35 a
2700 s4.87 b3.22 b0.20 a0.34 a
† Means within a column with different lower-case letters are significantly different at p < 0.05.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Seuferling, C.; Brye, K.; Della Lunga, D.; Brye, J.; Daniels, M.; Wood, L.; Greub, K. Evaluation of Greenhouse Gas-Flux-Determination Models and Calculation in Southeast Arkansas Cotton Production. AgriEngineering 2025, 7, 213. https://doi.org/10.3390/agriengineering7070213

AMA Style

Seuferling C, Brye K, Della Lunga D, Brye J, Daniels M, Wood L, Greub K. Evaluation of Greenhouse Gas-Flux-Determination Models and Calculation in Southeast Arkansas Cotton Production. AgriEngineering. 2025; 7(7):213. https://doi.org/10.3390/agriengineering7070213

Chicago/Turabian Style

Seuferling, Cassandra, Kristofor Brye, Diego Della Lunga, Jonathan Brye, Michael Daniels, Lisa Wood, and Kelsey Greub. 2025. "Evaluation of Greenhouse Gas-Flux-Determination Models and Calculation in Southeast Arkansas Cotton Production" AgriEngineering 7, no. 7: 213. https://doi.org/10.3390/agriengineering7070213

APA Style

Seuferling, C., Brye, K., Della Lunga, D., Brye, J., Daniels, M., Wood, L., & Greub, K. (2025). Evaluation of Greenhouse Gas-Flux-Determination Models and Calculation in Southeast Arkansas Cotton Production. AgriEngineering, 7(7), 213. https://doi.org/10.3390/agriengineering7070213

Article Metrics

Back to TopTop