B11D-0388
Filling Holes in Regional Carbon Budgets: Predicting Peat Depth and Volume in a North- Temperate Lake District
Peat deposits are estimated to contain 1/6 of all terrestrial fixed carbon (C) globally, and C in peat far exceeds that in live vegetation in many north-temperate and boreal landscapes. Because surface peat is more vulnerable to oxidation than deeply-buried peat, knowledge of peat depth distribution is required in order to judge the vulnerability of a region's peat C stores. However, regional estimates of C stored in peatlands are typically uncertain, largely because variation in peat depth is not well understood. To estimate the amount and spatial distribution of peat C in the Northern Highlands Lake District (NHLD) in northern Wisconsin, which contains 20% peatlands by area, we sampled 21 peatlands during summer 2008. Our study addressed two questions: (1) How spatially variable are peat depth and volume within and among peatlands of the NHLD? (2) To what degree can peat depth and volume be predicted from available spatial and/or field data? In each peatland (area range 0.4 to 24 ha), peat depth was measured on a regular grid, and interpolated to calculate total peatland volume and mean peat depth. Among the 21 peatlands, mean peat depth ranged from 0.2 to 5.1 m, with an average of 1.9 m, while volume varied by 3 orders of magnitude. Peat depth varied more within than among peatlands, and the maximum measured depth was >15 m. Mean and maximum peat depth could be predicted from local slope at the peatland-upland interface, measured either in the field or using digital elevation (DEM) data. Strikingly, field measurements (water chemistry, water table depth, vegetation cover) failed to substantially improve slope-based models. As the DEM data are widely available, this technique has the potential to considerably improve regional estimates of C stored in peatlands.
B11D-0389
Linking Carbon Balance to Spatial Patterns of Ground Subsidence in an Upland Tundra Ecosystem Affected by Climate Change
Global changes in atmospheric CO2 concentrations are causing temperatures to increase in northern high latitude ecosystems. A defining characteristic of these ecosystems is the presence of permafrost (perennially frozen ground), which currently maintain larges stores of organic carbon that are vulnerable to changes in temperature and precipitation. As global warming continues, vast areas of permafrost are predicted to thaw and large quantities of carbon could potentially be released to the atmosphere. This study is attempting to link changes in carbon balance to the physical changes that occur as permafrost thaws in an upland tundra ecosystem. We are using eddy covariance to attain both seasonal and annual C balance of the landscape. On average, daily net ecosystem exchange was -0.78 gCm-2d-1 during the summer growing season, but ranged from being a strong sink (-1.5 gCm-2d-1) to a C source (0.79 gCm-2d-1) depending on the day. Overall, the ecosystem was a C sink that accumulated 22.8, 25.3, and 21.7 gCm-2 in the months of June, July, and August, respectively. To gain a spatially explicit understanding of the contributing factors to C exchange we are combining our flux measurements with a digital elevation model of the source area surrounding the EC tower. Although our measurements to date show the ecosystem as a C sink during the short growing season, previous chamber measurements have shown that areas of extensive ground subsidence are the largest contributors to C flux, and can cause the ecosystem to become a C source over an annual basis.
B11D-0390
Boreal Forest Organic Soil Properties: Variation Within Soil Profiles and Across Landscapes
Organic soils play an important role in boreal ecosystem function by influencing temperature, hydrology, decomposition rates, fire dynamics, and species composition. Understanding the variability in soil properties both across a landscape and within a soil profile is very important given that relatively few soil descriptions exist in this region. Modeling studies, in particular, require a better understanding of the inherent variability in the boreal landscape. To help characterize organic soils of the boreal region we collected over 250 soil cores from mature and young black spruce (Picea mariana) stands located in both wet and dry ecosystems and stratified the soils by horizon type (moss, fibric, mesic, humic horizons). We investigated the porosity, bulk density, and carbon fraction of each horizon within the organic soil profile. Our results suggest that these properties vary by both horizon type and drainage class. We also found that a simple model including horizon type directly above the mineral soil and organic soil thickness can be used to accurately predict C stocks of organic soil profile. This relationship can be used to estimate C stocks (g/m2) for sites where horizons were described but sampling efforts did not occur, allowing more spatial representation of C storage across the boreal forest. We also present relationships between C density (g/m3) and height above mineral soil for both the dry and wet drainages, so that these values can be used in modeling studies of organic layer dynamics.
B11D-0391
Modeling interactions of soil hydrological dynamics and soil thermal and permafrost dynamics and their effects on carbon cycling in northern high latitudes
Large areas of northern high latitude ecosystems are underlain with permafrost. The warming temperature and fires deteriorate the stability of those permafrost, altering hydrological cycle, and consequently soil temperature and active layer depth. These changes will determine the fate of large carbon pools in soils and permafrost over the region. We developed a modeling framework of hydrology, permafrost, and biogeochemical dynamics based on our existing modules of these components. The framework was incorporated with a new snow dynamics module and the effects of soil moisture on soil thermal properties. The framework was tested for tundra and boreal forest ecosystems at field sites with respect to soil thermal and hydrological regimes in Alaska and was then applied to the whole Alaskan ecosystems for the period of 1923-2000 at a daily time step. Our two sets of simulations with and without considering soil moisture effects indicated that the soil temperature profile and active layer depth between two simulations are significant different. The differences of soil thermal regime would expect to result in different carbon dynamics. Next, we will verify the framework with the observed data of soil moisture and soil temperature at poor-drain, moderate-drain, and well-drain boreal forest sites in Alaska. With the verified framework, we will evaluate the effects of interactions of soil thermal and hydrological dynamics on carbon dynamics for the whole northern high latitudes.
B11D-0392
Soil Organic Carbon Residence Time in Japanese Temperate Forest: Insight From Radiocarbon Analysis of Density Fractionated Soil
Soil organic matter is an important carbon reservoir managing CO2 concentration in the atmosphere. However, the mechanism of soil carbon stabilization is little known. To explore the rate of C cycling in a temperate forest soil in Japan, at one of AsiaFlux monitoring sites, we sequentially density fractioned at 1.0, 1.6, 1.8, 2.1 and 2.4 g cm-3 on a soil at surface and deep soil layer and estimated turnover time of these SOC fractions. According to previous study in this site (Uchida et al., in prep.), large amount of light fraction (> 2.1 g cm-3) accounted at 35 - 50 cm depth and its age was significantly old as well as heavy fraction (> 2.1 g cm-3), although light fraction seems to consist of labile carbon. In this study, we investigated more better separation for labile fraction from mineral fraction in volcanic ash soil using different densities ranging from 1.6 to 1.8 g cm-3, excluding the most mineral and organio-mineral material from the light fraction. In surface layer (5 - 10 cm), carbon in lighter fractions (less than 1.8 g cm- 3) accounted for 42% of the total SOC their turnover times were shorter (6 - 43 yrs) than that in 1.8 – 2.1 g cm-3 (150 yrs). While, in deep layer (40 - 45 cm), lighter fraction comprised only small portion of total SOC (1%) but its age was significantly old (2038 yrs BP) as well as 1.8 - 2.1 g cm-3 (2335 yrs BP). The results partially support the previous study that volcanish-based Japanese temperate forest soil might be sequestrating carbon as light fractions semi-permanently.
B11D-0393
Ecosystem carbon balance and vulnerability of soil carbon in a drained lower coastal plain loblolly pine plantation
Coastal plain ecosystems comprise only about 5% of total U.S. land area, but the soil carbon density in these ecosystems is about 10-fold higher than in upland ecosystems and they may therefore play a disproportionately large role in ecosystem-climate feedbacks. The role of these ecosystems in continental carbon exchange is largely unclear because they have been underrepresented in flux monitoring networks. We monitored ecosystem carbon fluxes and pools for three years in two lower coastal plain loblolly pine plantations (3 and 17 years of age). The contribution of soil to ecosystem respiration decreased from over 90% immediately following a harvest to about 50% by age 17. The replenishment of soil C through litterfall exceeded heterotrophic respiration (Rh) by 2-9% in two years, but was 30% lower than Rh in the third year, highlighting the vulnerability of soil carbon stocks to interannual climate variability.
B11D-0394
A Meta-analysis of Timber Harvest and Site Preparation Effects on Soil Carbon Storage
Management practices can dramatically alter soil carbon (C) storage in forests. Timber harvesting and site preparation are a widely employed and studied form of forest management, yet abundant experimental data from this area of research have not recently been synthesized. We are using meta-analysis to test a database developed from 86 studies with published soil C storage values for paired harvested and un- harvested forests, in order to identify how timber harvesting and site preparation affect soil C pool sizes. Most of the studies in the database are from coniferous or hardwood forests of the continental United States, although temperate forests of Asia, Australia, Canada, and Europe also are represented. We have identified factors that influence soil C responses to harvest at global to regional scales, and estimated soil C storage shifts in pools of different vulnerability. At the global scale, soil C storage changes due to harvest differ according to soil horizon, soil taxonomic order, and species composition. Within soil types and at regional scales, climate, species composition, and harvest and site preparation methods appear to have more significant effects on forest soil C storage. At all spatial scales, forest floors and surface mineral soils show different levels of vulnerability to C loss or increase, highlighting the importance of constraining turnover times for C incorporated into these two soil pools as efforts to model the C cycle improve. As part of a larger effort to understand how soil C pools are impacted by management and global change, our meta-analysis identifies opportunities for increased soil C storage, situations where soil C losses are highly probable, and areas requiring improved understanding of mechanisms of forest soil C accumulation and loss.
B11D-0395
Simulated In Situ Determination of Soil Profile Organic and Inorganic Carbon With LIBS and VisNIR
There is growing need for rapid, accurate, and inexpensive methods to measure, and verify soil organic carbon (SOC) change for national greenhouse gas accounting and the development of a soil carbon trading market. Laser Induced Breakdown Spectroscopy (LIBS) and Visible and Near Infrared Spectroscopy (VisNIR) are complementary analytical techniques that have the potential to fill that need. The LIBS method provides precise elemental analysis of soils, but generally cannot distinguish between organic C and inorganic C. VisNIR has been established as a viable technique for measuring soil properties including SOC and inorganic carbon (IC). As part of the Big Sky Carbon Sequestration Regional Partnership, 240 intact core samples (3.8 x 50 cm) have been collected from six agricultural fields in north central Montana, USA. Each of these core samples were probed concurrently with LIBS and VisNIR at 2.5, 7.5, 12.5, 17.5, 22.5, 27.5, 35 and 45 cm (+/- 1.5 cm) depths. VisNIR measurements were taken using an Analytical Spectral Devices (ASD, Boulder, CO, USA) Agrispec spectrometer to determine the partition of SOC vs. IC in the samples. The LIBS scans were collected with the LANL LIBS Core Scanner Instrument which collected the entire 200 - 900 nm plasma emission including the 247.8 nm carbon emission line. This instrument also collected the emission from the elements typically found in inorganic carbon (Ca and Mg) and organic carbon (H, O, and N). Subsamples of soil (~ 4 g) were taken from interrogation points for laboratory determination of SOC and IC. Using this analytical data, we constructed several full spectrum multivariate VisNIR/LIBS calibration models for SOC and IC. These models were then applied to independent validation cores for model evaluation.
B11D-0396
The influence of agricultural management on soil's CO2 regime in semi-arid and arid regions
Two of the more important parameters which may help us better evaluate the impact of agricultural practices on the global carbon cycle are the in-situ soil pCO2 profile and the corresponding CO2 fluxes to the atmosphere. In an ongoing study, we monitored the pCO2 to a depth of 5 m in two adjacent irrigated Avocado orchards in the coastal plain of Israel (semi-arid region), and to a depth of 2 m in a semi- arid rain-fed and a arid rain-fed wheat fields in southern Israel. The soil pCO2 profiles and CO2 fluxes measurements were supplemented by measurements of soil moisture and temperature. The results showed differences in the CO2 profiles (both in the depth of the highest concentration and its absolute values) and the CO2 fluxes between the orchards and the wheat fields as well as along the year. In the irrigated Avocado orchards pCO2 values were in the range of 1.5 kPa at a depth of 0.5 m up to 8 kPa at depths of 3-5 m (even though Avocado trees are characterized by shallow roots). Such levels could affect reactions (e.g., enhancement of inorganic carbon dissolution) that may take place in the soil and some of its chemical properties (e.g., pH). As expected, soil pCO2 was affected by soil moisture and temperature, and the distance from the trees. Maximum soil respiration was observed during the summer when the orchards are under irrigation. In the wheat fields pCO2 level ranged from 0.2- 0.6 kPa at a depth of 0.2 m to 0.2-1 kPa at depths of 1-1.5 m (in arid and semiarid respectively). These pCO2 levels were much lower than those obtained in the irrigated orchards and seemed to depend on the wheat growing cycle (high concentration were noted at depth of 1-1.5 m close to the end of grain filling) and precipitation gradient (arid vs. semiarid). Since CO2 fluxes are directly affected by the pCO2 profile and soil moister and temperature the CO2 fluxes from the wheat fields were much lower (0.02- 0.2 ml min-1 m-2) compared to those obtained from the Avocado orchards (2-7 ml min-1 m-2). Our results clearly demonstrate the large variability in soil pCO2 concentration and flux to the atmosphere, and its dependence on the soil moisture regime (annual precipitation and irrigation) and type of cropping (orchard vs. field crop).
B11D-0397
Soil carbon in savanna landscapes – spatial pattern, uncertainty, and scaling
Woody plant invasion into grasslands and savannas has significant impacts on soil organic carbon (SOC) storage and its spatial heterogeneity. However, our understanding of spatial heterogeneity and uncertainty of SOC and its relationship to spatial patterns of vegetation in savanna landscapes remains limited. This understanding is essential for effective assessment and monitoring of SOC storage, turnover, and vulnerability in savanna landscapes. In this study, we investigated the spatial pattern of SOC and its relationship to that of vegetation patterns in a subtropical savanna in south Texas using spatially-explicit intensive sampling and spatial statistical analysis. We found that the spatial distribution of SOC was closely related to the spatial distribution of woody vegetation, and that there were strong within-patch patterns related to past dynamics of the woody vegetation. Results of conditional stochastic simulations showed significantly greater levels of uncertainty of SOC estimations in larger woody patches than in smaller woody patches and grassland, likely caused by complex canopy structure, root distribution and animal disturbance. Assessment of alternative sampling designs demonstrated the effect of spatial uncertainty on estimation accuracy of SOC storage, and helped generate effective sampling strategies to improve SOC estimation accuracy. This understanding of spatial uncertainty of SOC enabled improved approaches to estimate and monitor soil carbon storage over large landscapes based on remote sensing.
B11D-0398
Carbon cycling in fine roots of several mature forests: results using either locally-derived or bomb-derived radiocarbon enrichment
This work seeks to improve our ability to quantify C cycling rates in fine roots of trees in mature deciduous and coniferous forests. We use two different types of atmospheric 14CO2 enrichment to trace the time elapsed since C in plant tissues was fixed from the atmosphere by photosynthesis. The first uses a local enrichment of 14CO2 which occurred in early summer 1999, at the Oak Ridge Reservation, Tennessee. The second, employed at three different sites, uses the global enrichment in background atmospheric 14CO2 caused by thermonuclear weapons testing (bomb-14C). In both cases we employ a new model (Radix1.0) to track C and 14C fluxes through fine root populations. Radix simulates two live-root populations (the longer-lived one having structural and non-structural C components), two dead-root pools, non-normally distributed root mortality turnover times, a stored C pool, seasonal growth and respiration patterns, a best-fit to measurements approach to estimate model parameters, and Monte Carlo uncertainty analysis. Our results show that: (1) New fine-root growth contains a lot of stored C (~55%) but it is young in age (0.7 y). (2) The effect of stored reserves on estimated ages of fine roots is unlikely to be large in most natural abundance isotope studies. However, models should take stored reserves into account, particularly for pulse labeling studies and fast-cycling roots (< 1 y). (3) Radiocarbon values show a stronger correlation with position on the root branch system than they do with diameter or depth in the soil profile. (4) Live fine root dynamics are well described by a short-lived and a long-lived population, with mean turnover times <1 y and ~12 y, respectively. (5) Dead root decomposition is best modeled with (at least) two pools, with moderate (~2 y) and slow (~10 y) decomposition turnover times. (6) Root respiration has a large effect on fine root biomass and isotopic composition, and should be included in ecosystem C and isotope models. (7) It is important to distinguish structural from non-structural components in the long-lived root pool. Otherwise the 14C signature of root respiration is significantly different than atmospheric. We conclude that realistic quantification of C flows through fine roots requires a model with a level of complexity similar to Radix. Moreover, future root research efforts should seek to sample and sort roots by position on the root branch system rather than by diameter size class and improve estimates of root respiration within fine root populations.
B11D-0399
Universal Distribution of Litter Decay Rates
Degradation of litter is the result of many physical, chemical and biological processes. The high variability of these processes likely accounts for the progressive slowdown of decay with litter age. This age dependence is commonly thought to result from the superposition of processes with different decay rates k. Here we assume an underlying continuous yet unknown distribution p(k) of decay rates [1]. To seek its form, we analyze the mass-time history of 70 LIDET [2] litter data sets obtained under widely varying conditions. We construct a regularized inversion procedure to find the best fitting distribution p(k) with the least degrees of freedom. We find that the resulting p(k) is universally consistent with a lognormal distribution, i.e.~a Gaussian distribution of log k, characterized by a dataset-dependent mean and variance of log k. This result is supported by a recurring observation that microbial populations on leaves are log-normally distributed [3]. Simple biological processes cause the frequent appearance of the log-normal distribution in ecology [4]. Environmental factors, such as soil nitrate, soil aggregate size, soil hydraulic conductivity, total soil nitrogen, soil denitrification, soil respiration have been all observed to be log-normally distributed [5]. Litter degradation rates depend on many coupled, multiplicative factors, which provides a fundamental basis for the lognormal distribution. Using this insight, we systematically estimated the mean and variance of log k for 512 data sets from the LIDET study. We find the mean strongly correlates with temperature and precipitation, while the variance appears to be uncorrelated with main environmental factors and is thus likely more correlated with chemical composition and/or ecology. Results indicate the possibility that the distribution in rates reflects, at least in part, the distribution of microbial niches. [1] B. P. Boudreau, B.~R. Ruddick, American Journal of Science,291, 507, (1991). [2] M. Harmon, Forest Science Data Bank: TD023 [Database]. LTER Intersite Fine Litter Decomposition Experiment (LIDET): Long-Term Ecological Research, (2007). [3] G.~A. Beattie, S.~E. Lindow, Phytopathology 89, 353 (1999). [4] R.~A. May, Ecology and Evolution of Communities/, A pattern of Species Abundance and Diversity, 81 (1975). [5] T.~B. Parkin, J.~A. Robinson, Advances in Soil Science 20, Analysis of Lognormal Data, 194 (1992).
B11D-0400
Nitrogen Additions and Microbial Biomass: A Global Meta-analysis
Nitrogen (N) enrichment is an element of global change that could influence the growth and abundance of many organisms. In this meta-analysis, I synthesized responses of microbial biomass to N additions in 82 published field studies. I hypothesized that the biomass of fungi, bacteria, or the microbial community as a whole would be altered under N additions. I also predicted that changes in biomass would parallel changes in soil CO2 emissions. Microbial biomass declined 15% on average under N fertilization, but fungi and bacteria were not significantly altered in studies that examined each group separately. Moreover, declines in abundance of microbes and fungi were more evident in studies of longer durations and with higher total amounts of N added. In addition, responses of microbial biomass to N fertilization were significantly correlated with responses of soil CO2 emissions. There were no significant effects of biomes, fertilizer types, ambient N deposition rates, or methods of measuring biomass. Altogether, these results suggest that N enrichment could reduce microbial biomass in many ecosystems, with corresponding declines in soil CO2 emissions.
B11D-0401
Linear Theory of Soil Organic Carbon Dynamics: Implications in Modeling Soil Respiration and Carbon Sequestration
The long-term, large-scale soil organic carbon dynamics are typically described by mathematical models based on networks of linear reservoirs. Properties of these networks can be diagnosed from linear system theory (i.e. impulse-response transformations), which is seldom used in soil biogeochemistry, although it can be used to compare and test different models in the context of long-term carbon sequestration in soils. In this work, the general theory of linear impulse-response systems is briefly reviewed and linked to the theory of stochastic point processes. Two characteristic times are considered, the residence time (i.e., the time spent by a molecule in the system) and age (the time elapsed since the molecule entered the system). Both are represented through their probability density functions, which are computed explicitly as a function of model structure. Different cases are analyzed and compared, ranging from a simple individual-pool model, to feedback models involving loops (as in models of soil organic carbon-microbial interactions and physical adsorption-desorption), and to more complex networks often used to simulate in the details the soil organic carbon processes. As examples for these complex networks, the compartmental model CENTURY (Parton et al., 1987), and the continuum-quality Q-model (Agren and Bosatta, 1996) are considered. We assess the relative importance of model structural characteristics to determine the organic carbon residence time and age distributions.
B11D-0402
A new technique for continuous long-term measurement of CO2 efflux
Monitoring of soil CO2 efflux is an important tool in establishing ecosystem carbon balance. For many remote systems there has been a limited amount of robust observational data collected. This is particularly true at high latitudes where climate warming has the potential to accelerate decomposition of soil carbon currently locked up in permafrost. Recent research suggests a significant future reduction in Arctic permafrost, containing anywhere from 20 to 60 percent of the world's soil carbon. An accelerated decomposition of this carbon reservoir could induce a critical positive climate feedback and further increase greenhouse gas concentrations, and presumably lead to an acceleration of the climate system warming. Unfortunately, there is little monitoring done at high latitudes due to accessibility, costs, and limitations of existing measurement tools. We have been developing an inexpensive new instrumentation platform designed for long-term, unattended deployments in remote and challenging environments, which may aid efforts to understand climate-feedback related processes at remote locations. The instrument measures CO2 efflux continuously, and at high frequency (typically one minute or less), has few or no moving parts depending on embodiment, and is based on a new mathematical approach to soil gas transport and emission. The probe's simplicity and low cost of production allows for network-style deployments (many probes across a landscape) and/or to complement other measurement infrastructure. Here we present technical details of the methodology, characteristic data from lab and field trials, and priorities for future R&D. The long-term objective of this work is to establish a reliable technology that can be the basis for a soil CO2 efflux monitoring network with nodes across the North.
B11D-0403
Assessing NIR & MIR Spectral Analysis as a Method for Soil C Estimation Across a Network of Sampling Sites
Monitoring soil C stocks is critical to assess the impact of future climate and land use change on carbon sinks and sources in agricultural lands. A benchmark network for soil carbon monitoring of stock changes is being designed for US agricultural lands with 3000-5000 sites anticipated and re-sampling on a 5- to10-year basis. Approximately 1000 sites would be sampled per year producing around 15,000 soil samples to be processed for total, organic, and inorganic carbon, as well as bulk density and nitrogen. Laboratory processing of soil samples is cost and time intensive, therefore we are testing the efficacy of using near-infrared (NIR) and mid-infrared (MIR) spectral methods for estimating soil carbon. As part of an initial implementation of national soil carbon monitoring, we collected over 1800 soil samples from 45 cropland sites in the mid-continental region of the U.S. Samples were processed using standard laboratory methods to determine the variables above. Carbon and nitrogen were determined by dry combustion and inorganic carbon was estimated with an acid-pressure test. 600 samples are being scanned using a bench- top NIR reflectance spectrometer (30 g of 2 mm oven-dried soil and 30 g of 8 mm air-dried soil) and 500 samples using a MIR Fourier-Transform Infrared Spectrometer (FTIR) with a DRIFT reflectance accessory (0.2 g oven-dried ground soil). Lab-measured carbon will be compared to spectrally-estimated carbon contents using Partial Least Squares (PLS) multivariate statistical approach. PLS attempts to develop a soil C predictive model that can then be used to estimate C in soil samples not lab-processed. The spectral analysis of soil samples either whole or partially processed can potentially save both funding resources and time to process samples. This is particularly relevant for the implementation of a national monitoring network for soil carbon. This poster will discuss our methods, initial results and potential for using NIR and MIR spectral approaches to either replace or augment traditional lab-based carbon analyses of soils.
B11D-0404
Simulating CO2 Released from Soil: A Bayesian Approach
Measuring CO2 efflux in soil is important in understanding soil carbon dynamics. Current approach in simulating CO2 release employs different types of kinetic models, and least square algorithms are often used. The most serious problem exists in the current method is the uncertainty in error distribution of cumulative CO2 release data. Since Bayesian approach uses priors to update original replicated data to obtain a posterior distribution of estimated parameters, we use the Bayesian statistical approach as an alternative for model simulation to circumvent this uncertainty of error distribution in cumulative CO2 release data. Comparing the results of Bayesian and frequentist statistics among four models, i.e. single exponential, single exponential and linear, double exponential, and second order kinetics, both frequentist and Bayesian statistics lead to selection of the double exponential as the best fit for CO2 release from three organic substrates (fish byproducts) applied to soil. The estimated parameters were similar between frequentist and Bayesian statistical results, and there was small difference in curve fitting of cumulative CO2 release data. Because Bayesian statistics avoided uncertainty in error distribution of CO2 release data, it can serve as an alternative for such kind of simulation before an adequate frequentist statistical method(s) is developed to handle such uncertainty of error distribution of cumulative CO2 release data.