Peatland Stratigraphy as a Proxy for Long-Term Carbon Dynamics: A Case Study from Estonia

Abstract

Sustainable management of peatlands is one of the key global strategies for mitigating climate change. The balance between carbon (C) sequestration and emission in peatlands reflects environmental conditions over time and can provide insight into long-term ecosystem dynamics. However, current methods for estimating greenhouse gas (GHG) fluxes are often labor-intensive, costly, and site-specific. In this study, we propose a simplified and cost-efficient method to estimate long-term carbon balance in peatlands based on the inorganic (mineral) content of drill core samples. The approach uses exponential decay equations to approximate peat accumulation and decomposition processes over time. A conceptual model is applied that accounts for both anaerobic transformation of organic matter of varying molecular complexity and enhanced aerobic decomposition resulting from anthropogenic drainage during the last century. The model was applied to more than 100 drill cores from four peatland systems in Estonia. The resulting trends were compared qualitatively with known climatic fluctuations of the last millennium, including periods associated with the Little Ice Age. The results suggest that, in many cases, carbon losses from decomposition in deeper peat layers may exceed carbon accumulation in upper layers, even in peatlands that appear to be well preserved. The proposed method provides a rapid, low-cost, first-order approximation of peatland carbon dynamics and may serve as a complementary tool for large-scale assessments where detailed process-based models are not feasible.

1. Introduction

Peatlands play a crucial role in the global carbon cycle, acting both as carbon sinks and sources depending on environmental conditions. The balance between carbon sequestration and emission is controlled by hydrology, vegetation, climate, and anthropogenic impacts. These processes are recorded in the vertical structure of peat deposits, which can therefore be interpreted as archives of past environmental conditions.

Peatlands cover only 3% of the Earth’s total land area but store 30% of the world’s soil carbon (C) [1]. Moreover, 15% of all peatlands worldwide have been drained [2] for use in agriculture, grazing, forestry (often for bioenergy plantations), and peat mining. These drained peatlands are responsible for almost 6% of anthropogenic carbon dioxide (CO₂) emissions when peat fires are included [3], representing nearly a quarter of emissions from the entire land use, land use change, and forestry (LULUCF) sector.

In Estonia, 22% of the whole territory is covered by peatlands. Nearly 70% of Estonian peatlands have been affected by drainage works undertaken during the 20th century to yield land for use in agriculture and forestry. In these areas, the peat accumulation has ceased, and mineralization of organic matter has replaced the C accumulation [2,4]. Based on field inventories in 2009 and 2010, preserved mires formed at least 240,000–245,000 ha or ca 5.5% of Estonian territory, which was 2.6–2.8 times less than 60 years ago.

Intensive peatland management increases emissions of CO₂, methane (CH₄), and nitrous oxide (N₂O), the main GHGs associated with wetland environments [5]. The loss of peat C stock can be considered virtually irreversible, since its recovery takes place over a timeframe of thousands of years. Peatland conservation, restoration, and improved management are important instruments in the mitigation of climate change, providing considerable mitigation with relatively low effort. There is therefore a need for relatively simple yet realistic C exchange models between the atmosphere and peatlands [6].

Monitoring peatlands and protecting them from drought and artificial drainage is essential to preserve their sequestered C. Many attempts have been made to estimate the CO₂ sequestration or emission history and assess the C stock of peatlands with different methods. The most common and well-known method is radiocarbon ¹⁴C dating [7,8], which is based on the constant rate of production of ¹⁴C in the atmosphere, incorporation of radiocarbon into living plants and its subsequent decay. Radiocarbon dating permits the construction of timescales for Late Pleistocene and Holocene records spanning the last ca 50 ka, but it requires using expensive equipment and is labor-intensive.

In situ measurement of the CO₂ flux can be used to characterize the present peat decomposition speed (a fraction of emitted CO₂ also originates from respiration of living roots) and can be measured directly [9], but requires a measurement period of at least 1 year and can provide only data about the current status of the peatland. Several specific methods have been described, for example, dating peat layers using ²¹⁰Pb chronology [2,10,11,12], detecting spheroid anthropogenic carbonaceous particles in peat layers [11], and tephrochronology (based on the detection, extraction, and analysis of volcanic ejecta) [13,14].

GHG emissions from the agricultural and LULUCF sectors are often underestimated or even ignored. In some areas, such as the subarctic zone, GHG emissions from these sectors exceed all other emissions (including from the energy and transportation sectors) [15], and are mostly related to drained peatlands. This drainage and the resulting groundwater level reduction have switched C sinks to C sources in many peatlands, profoundly altering their role in global ecosystems. As a result, the assessment of peatland C balances is crucial, but, to date, no simple method applicable for large areas is available. Recent modeling of the atmospheric CO₂ drawdown rate (defined as the CO₂ decrease in ppm over the number of days in spring or summer) between 1974 and 2014 in the northern high latitudes (north of 50° N) suggests that warmer springs have reduced C uptake. Similarly, warmer summers have led to a greater C release [16]. For example, the C accumulation rate during the Holocene thermal maximum was nearly six-fold the current rates observed at peatland sites in Minnesota, Alaska, and Canada [17].

Carbon losses from drained peatlands have been estimated indirectly, but these methods are often complex or based on approximate and indirect data [18]. For example, Wüst-Galley et al. [19] used a combination of comparing historic soil maps, wetland toponyms, and literature data. Krüger et al. [20] studied vertical depth profiles of isotopic ratios such as the ¹³C-to-¹²C ratio and the ¹⁵N-to-¹⁴N (nitrogen) ratio, as well as ash content, carbon-to-nitrogen ratio, and bulk density alongside radiocarbon dating to identify peat degradation and calculate C loss. Webster et al. [21] used the process-based Wetland-DNDC model (i.e., DeNitrification-DeComposition), which was parameterized, calibrated, and corroborated by field data to examine boreal fen peatlands along a nutrient gradient in northern Ontario, correlating annual variability in C effluxes to meteorological conditions and water table. DNDC-type models simulate C and N biogeochemistry in natural and agro-ecosystems and can be used for predicting emissions of CO₂, CH₄ and trace gases. The Wetland-DNDC model integrates forest growth, hydrology, and soil biogeochemistry modules and uses data on climate, water table, vegetation type (for tree, shrub, moss, and herb layers), age and biomass, as well as plant physiological parameters and soil properties as inputs [21]. The real-time emissions can also be estimated using geophysical methods such as linking spatial frequency-domain electromagnetic induction measurements to respiration of CO₂ [22].

The described models are: complex; require extensive, long-term, and costly field and laboratory work; and do not allow direct estimation of past emissions. Few models have been proposed to quantify the past C release during the Holocene.

Frolking et al. [6] developed a Holocene Peat Model (HPM) that quantifies annual net primary productivity based on peat and water table depth. In the HPM, the annual C accumulation is calculated as the net balance of primary productivity (both above and below ground) and litter and peat decomposition. The latter is dependent on the estimated water table in the peat column, which controls the degree of anoxia and the depth to the anoxic zone. Peat bulk density is expressed as a function of the humification degree. The outputs of HPM include both time series of annual net C and water fluxes as well as peat accumulation. Spahni et al. [23] proposed a similar peatland module embedded in a dynamic global vegetation model (Land surface Processes and eXchanges, LPX) and featuring functional types of peatland vegetation, moisture- and temperature-dependent respiration rates, dynamic C exchange between upper oxic and deep anoxic layers (henceforth acrotelm and catotelm, respectively), and a dynamic nitrogen (N) cycle. Wang et al. [17] developed a more complex model that integrated the dynamics of hydrology, soil thermal regime, CH₄ formation, and peatland C and N contents in a simulation of long-term peat C accumulation in Alaska during the Holocene. The model was applied for a 10,000-year period.

All of the above-described models are quite complicated and require multiple factors to be taken into consideration that cannot be directly ascertained or measured. No reputed simple integral models based on exponential decay kinetics equations have yet been proposed for peatlands.

The aim of this study was to create a simple and cost-efficient model to estimate historic C sequestration and emissions of peatlands in Estonia, which is a very suitable area for modeling these processes. More than ¾ of wetlands are strongly affected by anthropogenic disturbances, which have occurred within a very short timeframe (ca last 100 years). Almost all major peatlands in Estonia are well studied by drilling, and the main data of drill core chemical and botanical contents are freely accessible through the peatlands database [24]. However, the radiocarbon calibration data are scarce.

Because of this, we were required to take a quite simple approach based on molecular structure and molecular weight, which is preferable for mathematical modeling compared with complex models that thoroughly consider the hydrology of bogs and the functional groups of vegetation. Even without strong calibration, the integral values of sequestration/emission of carbon allow for estimating the wetlands’ health and excessive anthropogenic organic matter decay. The model can easily be applied to different types of peatlands in various climatic zones globally.

In addition, the development of simplified approaches for estimating peatland carbon balance is directly relevant for sustainable land-use planning and climate policy, particularly in regions where detailed process-based modeling is not feasible.

2. Materials and Methods

2.1. Data Sources

Data from most Estonian mires and peatlands (32,942 measurements from 558 peatlands) have been collected over several years, and these data are now available in a digitized form in the ODBL license open access database [24] developed by Tallinn Technical University (the abbreviation ODBL denotes Open Data Commons Open Database). Most of the data were systematized from manual reports and have been made publicly available [25,26].

In general, the sampling interval through the peat profile at each measurement point was 0.25 m, from which the botanical composition, decomposition rate, ash content, natural moisture, and acidity (pH) were analyzed. Peat properties were determined according to a uniform methodology [25]. All measurements were taken throughout the whole peat deposit, in most cases to the bedrock sediment.

Sampling campaigns were conducted using standardized field protocols. While seasonal variation may affect surface peat layers, deeper peat strata reflect long-term accumulation processes and are not significantly influenced by short-term environmental variability.

2.2. Study Area

For the application of the developed model, we selected four mires in south-western Estonia (Figure 1) that have been extensively researched by drilling. The study area emerged from the receding Littorina Sea approximately 4000 years ago during a very short period of time due to post-glacial rebound. A description of the studied sites is provided in

Table 1. Brief description of the study sites [27].

Study Site	Lavassaare	Tolkuse- Soometsa	Nigula-Tõrga-Rongu	Kikepera- Mustraba
Study area (ha)	21,868	5503 + 1631	2320 + 1291 + 1348	8725 + 2310
Size of excavation area (ha)	19,701.9	4129.32 + 1148	2078 + 829 + 770	6144 + 1450
Peat deposit types (ha)	Lavassaare:	Tolkuse:	Nigula:	Kikepera:
	Fen soils 7490 Mixed-raised bog 4794 Raised bog 9584	Fen bog 3472 Mixed raised bog 412 Bog 1619	Fen bog 2320 Mixed-raised bog 242 Raised bog 2078	Fen bog 3175 Mixed bog 1475 Raised bog 4075
		Soometsa:	Tõrga:	Mustraba:
		Fen bog 750 Bog 881	Fen bog 624 Raised bog 667	Fen bog 1472 Raised bog 838
			Rongu:
			Fen bog 578 Raised bog 770
Site description	Fed by rain and groundwater. Thin sapropel layer (0.1–0.2 m) under the peat deposit. Peat and sapropel are lying on loam, sandy loam and sand. The dominant bog type is open bog (50–60%). Peat has been industrially mined there for more than 100 years [27].	Bogs are formed from lagoon lakes between the dunes of the Littorina Sea and Ancylus Lake and fed by rainwater. The peat is lying on sand (glacial till in the southern part). Mostly forested.	Mainly rain-fed blanket bogs lying on glacial till. Open bogs with lakes and bog islands.	Raised bogs formed from lakes (lake mud layer 0.2–0.3 m under the peat deposit) and fed by rainwater. Mostly rusty bog moss (Sphagnum fuscum) deposit.
Soil	Fen, transitional bog and bog soils. Histic and Eutric Histosols.	Fen, transitional bog and bog soils. Histic and Eutric Histosols.	Fen, transitional bog and bog soils. Histic and Eutric Histosols.	Fen, mixed bog and bog soils. Histic and Eutric Histosols.
Plant association or dominant plant species	Lavassaare:	Tolkuse:	Nigula:	Kikepera:
	Fen type forest on the bog edges. Transitional bog Pinus sylvestris Betula sp. mixed forest containing several bog species (Ledum palustre L., Calluna vulgaris L., Oxycoccus palustris Pers., Scheuchzeria palustris).	Picea abies–Betula sp. forest types. Transitional bog forests with Ledum palustre, Vaccinium myrtillus L., Rubus chamaemorus L., Vaccinium vitis-idaea, Eriophorum vaginatum, Carex sp., Bryopsida sp. in undergrowth.	Treeless and forested ombrotrophic raised bog.	Fen type forest Treeless and treed ombrotrophic raised bogs on the area edges.
			Tõrga:	Mustraba:
			Open bog with a large number of Vaccinium oxycoccucos L. and Rubus chamaemorus L.	Treeless or treed ridge–hollow bog.
		Soometsa:	Rongu:
		Treeless and treed ombrotrophic raised bog with ridge-hollows.	Mixed bog types (treeless and treed ombrotrophic raised bog). Fens with sparse or no trees.
Depth of peat deposit (m)	Up to 9.9	1.1–6.8	4.25–7.0	1.5–7.45
Number of peat samples analyzed	754	121 + 101	58 + 103 + 84	269 + 164

together with the average depth of the profiles considered and the range of variation in depth among the considered sites. The study sites are listed according to the original names listed in the used database. Similar peat deposits regarding the location are listed together as composite sites (Tolkuse-Soometsa; Nigula-Tõrga-Rongu; Kikerpera-Mustraba).

Data for the uppermost stratum (within the root zone of living plants) was discarded as this zone contains live peat moss, live roots, and fragments of vegetation. Similarly, data from the lowermost stratum, where peat may be mixed with underlying mineral soil, were discarded.

2.3. Modeling Approach

A model of peatland C balance is based on equations that apply physical laws and assumptions to calculate the corresponding outputs based on local measurements of the emitted and accumulated gases. In the current study, a MATLAB-based integral model was developed.

As the input data, the organic content arrays from the peatlands database described above were used. The maximum measurement depth reached 9 m, and the measurement step was 25 cm. The time step for equations was one year, and the total time extent was ca. 4000 years (total age of the mires included in the research). The methodology developed in this study allows the C sequestration and emissions to be assessed both at present and throughout the peat bog lifetime based on chemical analysis of the drill core samples. The data obtained from a single drill core at every measurement point sufficiently characterizes the C behavior of the peatland at a given location.

We analyzed 101 drill cores from the four peatlands. The total thickness of the peat layer (length of the drill cores until reaching the subsoil layer of the peat) in meters is presented in Figure 2.

The average inorganic matter content and the same in the top strata is presented in Figure 3. As the very upper layer of peat contains live plants and moss, and the bottom layer is probably mixed with underlying inorganic soil, the first and last 25 cm of the cores were excluded from analysis.

The first successful attempt for the estimation of soil organic C based on first-order kinetics was proposed by Hénin and Dupuis in 1945 [28]. Later, modifications of the Hénin–Dupuis equation have been used to model the dynamics of soil organic matter in manured croplands, in which crop residues are returned to soil. Such modifications take into consideration different decay rates of organic matter from different amendments such as manure, crop roots, and returned crop residues [29].

In semi-anaerobic conditions with very limited availability of oxygen, the peatland C balance can be presented by a linear differential equation [28] describing the humification of soil organic matter:

dC_s/dt = h∙C_i − k∙C_s

(1)

where C_s = C_s(t) is the soil organic matter content (Mgha⁻¹), t is time (year), h is the humification constant (unitless), C_i is the annual C input (Mgha⁻¹ year⁻¹), and k is the apparent decomposition rate of specific organic fraction in the soil (year⁻¹).

The general solution of the corresponding homogeneous equation is as follows:

dC_s/dt = −k∙C_s

(2)

in which C_s(t) = a∙e^−k∙t, where a is an arbitrary constant depending on the initial conditions. Thus, the general solution of Equation (1) can be written as

C_s(t) = a∙e^−k∙t + y(t)

(3)

where y = y(t) is a particular solution of Equation (1).

With immediate replacement, it is easy to make sure that the function y = h∙C_i∙(1 − e^−k∙t)/k satisfies Equation (1). If the residue returns can be considered constant over the integration interval, the equation, similar to the two-compartment Hénin–Dupuis model with two pools of organic matter, is as follows:

C_s(t) = a∙e^−k∙t + h∙C_i∙(1 − e^−k∙t)/k

(4)

where the first component describes the degradation of biomolecules with lower molecular weight (e.g., peptides, fats, and sugars) and the second exponent describes the slow oxidation of humic substances. Similar to Bertora [29], we take into consideration different decay rates of different fractions of organic matter. The values of constants a, b, c, and d can be estimated from inorganic content values for each drill core, using MATLAB R2017a software. A 2 m core with fewer than 10 data points is not sufficient for building a serious time graph. However, in the case of peat thickness exceeding 3 m, the mathematical modeling can be provided.

We noticed that in the upper strata of peat deposits, the decomposition rate increased many times compared with the natural decomposition level. This can be explained by the anthropogenic lowering of the water level during the last century. The density of data does not allow exact modeling of the anthropogenic loss of organic matter because of the very short time period of anthropogenic influence; however, it allows us an opportunity to estimate the total loss of organic matter.

3. Results

In this study, we propose a simplified C sequestration/emission model based on the mineral content of peat layers acquired from the Estonian peatland database. These observations suggest that the method may provide proxy-type information that can support qualitative reconstruction of past climatic variability at a given location. Radiocarbon dating is unavailable for our study area; however, the timeline given from this method could be easily refined by linking drill core depth data to radiocarbon dating data.

We made a simplified assumption that all inorganic fraction in peat originates from groundwater and all supplied dissolved minerals are taken up by growing peat mosses and other vegetation. The increment in primary organic matter depends on meteorological conditions, but the amount of minerals (mainly Ca and Mg) supplied by groundwater remains unchanged, because the mineral content in groundwater depends on the equilibrium with bedrock (containing a fine fraction of limestone) [30].

Plotting the ash content of peat deposits against peat depth reveals two different types: typical (Figure 4) and atypical (Figure 5). On most graphs calculated, we observe a relatively linear increase in the mineral content of peat towards deeper layers, as well as a rapid decrease in mineral content in the upper layers.

Some graphs are atypical, containing sudden unlinearities, probably because of changes in the water regime during the peat formation process.

As the density of peat (dry matter) does not vary more than 5% in all cores, these differences were considered statistically insignificant. The following two main assumptions were taken:

Assumption 1.

The amount of organic matter annually deposited is generally stable, but fluctuations can occur due to variations in moisture, solar irradiation, and other external factors [31].

Assumption 2.

The rate of inorganic compounds in the deposited substance is mostly stable every year throughout this period.

These observations allow a tentative alignment of the time axis with known climatic events, although this comparison is qualitative and not used as a formal calibration.

Over time, some of the organic matter decomposes into CO₂ and CH₄ (in lower quantities) as well as water, but the mass of the mineral fraction remains unchanged. Thus, during the whole lifetime of the bog, T (in our example, 4000 years), the mass of mineral fraction accumulated per unit area is $m_{min}=\rho \sum s_n{k}_n$, where $s_n$ is the thickness of each layer (0.25 m in this study), $\rho$ is its average density and $k_n$ its mineral content. The amount of minerals deposited per year is equal to $m_{amin = {\displaystyle \raisebox{1ex}{$m_{min}$}\!\left/ \!\raisebox{-1ex}{$T$}\right.}}$.

Figure 6 shows the increases in the thickness of peat throughout the 4000-year period (an estimation of the age of each sample will be discussed later).

First, we approximated both functions by a single function, which fits the whole measurement series with the sum of two exponents:

$$F\left(t\right)=a\times e^{bt}+c\times e^{dt}$$

(5)

where the constants a, b, c, and d were estimated by the MATLAB model from the dataset. The results are shown in Figure 7. The green area (integral of the data function over all periods of peat deposition) represents the organic matter accumulated at the current point; the area to the left (red shading) is the amount of organic matter loss due to the decrease in water level.

The plots for the different measurement sites (six randomly chosen drill cores) obtained by using this method are similar in shape (Figure 8). The average ash content in the drill cores ranged from 2.3% ± 1.1% in the Nigula mire to 3.8% ± 2.1% in the Lavassaare mire. This represents a typical value for northern peatlands, which do not receive a large amount of dust from deflation and/or tephra ash inputs. The average value is lower than 10% in all Eurasian and Northern American regions [32] and is lower for humified peats.

In most graphs, three steep falls can be seen. Reconstructions of total solar irradiance show that these decreases occurred in approximately 1310, 1480 and 1680 (Figure 9). We can hypothesize that decreases in the organic substance accumulation rate correspond to the last millennium minima of solar irradiation. These assumptions allow us to first calibrate the axis of time (the timescales on Figure 8 were adjusted this way), and second, based on just drilling cores, the method may provide additional proxy-type information supporting the interpretation of past climatic variability at a given location.

4. Discussion

Decomposition of organic matter in peat is an extremely complex process. The process of decomposition of organic matter described by the first exponential component in function (2) finalizes in peatlands under observed conditions during ca 1000 years, and part of the deposited organic matter transforms slowly into humic substances, the extremely complex, large molecules (precursor for transformation of peat into coal). The second exponential component describes the very slow oxidation of humic substances. In the case of water level downturn (due to drainage), extremely fast oxidation of the organic material occurs upon exposure to atmospheric air.

The inorganic fraction in peat deposits may originate from different sources, but it is mainly derived from water. The mineral content of water assimilated by vegetation is at equilibrium with groundwater. As all wetlands investigated are located on limestone bedrock, the mineral content in it is quite stable [40,41]; however, flowing surface water may disturb this equilibrium. Air deposition of inorganic matter before the late 19th century can be considered negligible, as there are neither active volcanoes nor large plains where dust storms can form in the vicinity. Furthermore, the anthropogenic impact on the air deposition before the beginning of the Industrial Revolution in the Baltic region is insignificant [30].

The slow linear increase in the mineral content in the deeper layers can be explained by the slow, naturally occurring oxidation of organic matter in peat. The rapid increase towards the surface is caused by the decrease in water level due to the enhanced oxidation of organic matter after peat mining. Second, in the areas that have been drained in connection with peat mining, rapidly increasing mineralization is observed in some cases, associated with the exposure of the peat deposits to atmospheric oxygen. A similar pattern was found by Muller et al. [42] in geochemical and stratigraphic studies performed at Lynch’s Crater in Queensland, Australia. An increase in ash content in the topsoil was attributed either to the increased settling of atmospheric dust, enhanced decomposition of organic matter, or to a combination of both factors since the mid-Holocene. These changes were linked to a climate shift—wetter conditions associated with net accumulation of peat prevailed in this region during the early Holocene. Moreover, drought after the mid-Holocene, which has decreased the slowdown in peatland C accumulation during Holocene cooling events, has been described in many studies [6,17].

Atypical graphs, such as the one shown in Figure 5, have steep fluctuations, which can be explained by a sudden change in the conditions of peat formation, for example, a watercourse becoming overgrown with aquatic plants. These drilling cores are neglected in this study, as we lack the detailed floral biodiversity data required for a vertical profile of peat deposits in a layer-by-layer manner. This analysis is feasible but requires extensive pollen and radiocarbon studies and thus lies outside the scope of this work. However, even the simplified model allows evaluation of the dominant C flux of the given peatland (C sequestration or release).

As a basis of our simplified model, we make some assumptions, as we have no accurate justification for the peat layers’ exact age in this region (radiocarbon dating or similar). We presume that if the hydrological regime is not changed, the deposition of organic matter is dependent on climatic conditions, but if inorganic matter is nearly stable, it can be said to depend only on the available soluble mineral content of the groundwater [40]. If the hydrological regime has changed steeply, this can be easily ascertained using graph data, and the core is excluded from the dataset (see Figure 5).

In addition to drainage effects, changes in acrotelm thickness, vegetation dynamics, and increased atmospheric deposition in recent centuries may also contribute to observed mineral content variations.

When refining the model, it is easy to include the actual values of these parameters: the exact peat deposition rate can be obtained by the radiocarbon measurements across the entire profile and bulk density by the laboratory analysis of peat in the various layers. However, such data are currently unavailable for our study area.

Other recently published models also rely on first-order kinetics. Frolking et al. [6] modeled annual peat accumulation considering first-order decay of plant litter (for specific functional groups of plants) in the acrotelm and catotelm as a function of water table. Spahni et al. [23] modified a dynamic vegetation model in terms of hydrology- and temperature-dependent decay rates, the peatland N cycle, and the dynamic C exchange between acrotelm and catotelm. The model developed by Wang et al. [17] calculates the peatlands’ net carbon sequestration rate by subtracting the aerobic heterotrophic respiration of soil organic matter, CH₄ emissions, CO₂ emissions due to CH₄ oxidation, CO₂ emissions due to CH₄ production, and CO₂ emissions due to other anaerobic processes relating to net ecosystem production. All these models need a very large amount of input data. Our model, although less precise, is applicable in all cases when the peat layer has sufficient thickness, and it only requires chemical analysis of the drill core.

5. Conclusions

The developed model allows evaluating the condition of peatland C capture and/or release. In the graphs of carbon storing, we can see two integrals; the area represents the total amount of organic C captured in the given location (green area on Figure 7), and the uppermost area (pink area) represents organic C lost to anthropogenic activities (peatland draining). Thus, expensive and time-consuming direct measurements of CO₂ flux can be avoided, and modeling of the C balance of large areas (entire geographic regions) can be provided using simple and cheap drill core data. The developed methodology allows both C sequestration and emissions to be assessed for the present and throughout the peat bog lifetime based on analysis of the ash content of the drill core samples, as well as a potential trend extrapolation of C dynamics based on the fitted functions, although predictive use requires further validation. Overall, the data obtained from a single drill core, with minimal labor and time costs, can potentially characterize the condition of a peatland at a given location.

Using our methodology, it is possible to calculate C balances for large areas by just using drilling and geo-data, and this process can be fully automated.

A limitation of this study is the use of simplified assumptions, including relatively stable mineral input, limited consideration of bulk density variability, and the absence of radiocarbon-based age calibration. Therefore, the model should be interpreted as a first-order approximation rather than a fully mechanistic reconstruction. Future studies should integrate radiocarbon dating, pollen analysis, and hydrological data to improve temporal resolution and accuracy.

Further studies are needed, including more accurate dating and more frequent sampling of peat layers, using radiocarbon dating and pollen analysis to adjust the timescale. In the course of further investigation, additional factors such as water levels need to be taken into account. The model needs to be validated for other geographic regions and climatic conditions. Overall, our mathematical model allows rapid, simple, and cheap estimation of CO₂ behavior in large peatland areas. Using the same method, C models for forests and arable land could be developed. Finally, in the context of anthropogenic climate change, further development of this model provides the relevant institutions with information to support development plans for the sustainable management of peatlands.