Atmospheric Fourier Transform Infrared Monitoring of Ammonia and Ethylene near the Saint Petersburg Agglomeration (Russia)

Abstract

The atmospheric air quality is one of the crucial factors determining people’s health, duration and quality of life. The importance of ammonia (NH₃) and ethylene (C₂H₄) is due to the fact that they are precursors of secondary organic aerosols (SOA) and phytotoxicants, which significantly affect air quality, cause human diseases and damage plants. The Fourier Transform Infrared (FTIR) spectrometry is a powerful tool for long-term monitoring of the atmospheric gas composition, including toxic gases. The paper presents the results of atmospheric FTIR measurements of NH₃ and C₂H₄ at the St. Petersburg State University observational site (59.88° N, 29.83° E, 20 m above sea level) located in a suburb of greater Saint Petersburg. This work demonstrates the applicability of the ground-based atmospheric FTIR spectroscopy to long-term monitoring of air pollution in urbanized areas and in particular to provide information on the NH₃ and C₂H₄ abundance in the atmosphere, including the analysis of their annual cycle, long-term trends, and positive anomalies. It was shown that for NH₃ and C₂H₄, a statistically significant decrease in column-averaged dry-air mole fraction values (X_NH3 and X_C2H4) was observed, amounting to (−2.3 ± 0.2)%/year for the 2009–2025 period and with the rate (−2.2 ± 0.4)%/year for the 2016–2025 period, respectively. Periodically recorded X_NH3 anomalies indicate the presence of intensive emission sources in the region, subjecting ecosystems in adjacent areas to constant exposure to NH₃ concentrations exceeding the critical level. Anomalously high values of X_NH3 and X_C2H4 were recorded simultaneously only once—on 17 October 2017. Using data on HCN total column (as a forest fire indicator) and the results of atmospheric dispersion modeling, it was shown that this pollution event was caused by the influence of biomass burning products emitted from wildfires located approximately 250 km to the north-west from the observational site in the Helsinki area (Finland).

1. Introduction

Air pollution, along with its impact on public health, can produce a depressing effect on plant vegetation and photosynthesis processes. The class of phytotoxicants includes such trace gas components of the atmosphere as tropospheric ozone (O₃), sulfur dioxide (SO₂), ammonia (NH₃), ethylene (or ethene, C₂H₄), peroxyacetyl nitrate (PAN), hydrogen fluoride (HF), chlorine (Cl₂), etc. [1]. Their high concentrations, caused by powerful anthropogenic (industry, motor transport, agriculture) and natural (wildfires) emissions, can reduce the absorption of CO₂ by vegetation during photosynthesis and, under certain conditions, cause plant death.

The importance of this study is due to the fact that NH₃ and C₂H₄ are both phytotoxicants and precursors of secondary atmospheric aerosols (SOAs), which, in turn, have a significant impact on climate, air quality, and public health, and determine atmospheric visibility [1,2]. Ammonia is the most abundant alkaline compound of the Earth’s atmosphere [3]. Atmospheric NH₃ is responsible for a significant portion of the transport of reactive nitrogen over distances of up to hundreds of kilometers [4]. Changes in ammonia levels alter the biogeochemical nitrogen cycle by increasing the concentration of organic nitrogen and nitrite in the environment, which causes eutrophication and acidification of surface waters and soils. Ammonia can act as a phytotoxicant for some plants, making them less resilient to frost, drought, and other stressful conditions [3]. Pogány et al. [5] noted that modern global NH₃ emissions have increased by a factor of 2–5 compared to pre-industrial levels. Agriculture, especially livestock and crop production, is the primary source of ammonia emissions, accounting for 70–80% of total atmospheric NH₃. Industrial emissions, transport, solid and liquid waste management systems contribute less than 15% [3]. Natural sources of ammonia emissions include soil, vegetation, and wildlife, but these estimates are currently subject to large uncertainties [3]. Despite its short atmospheric lifetime (a day or less), NH₃ reacts with sulfuric (H₂SO₄) and nitric (HNO₃) acids to form a fine fraction of atmospheric aerosols (ammonium salts), which remain in the atmosphere for several days to a week [2]. Nevertheless, while ammonia does increase the mass of fine particles when it reacts with ambient acidic aerosols, it is also at the same time neutralizing acids and making the particulate matter less toxic. Thus, at atmospheric levels, NH₃ can act as protective of human health by lowering the toxicity of PM2.5: “…an excess of ammonia gave protection from levels of sulfuric acid which, in the absence of ammonia, would have caused 50% mortality” [6]. In recent years, due to improvements in measuring equipment and data processing approaches, there has been a growing interest in studying atmospheric NH₃, in particular using ground-based and satellite atmospheric observational systems [7,8,9,10].

Ethylene, being one of the most abundant unsaturated hydrocarbons in the atmosphere, is a plant hormone which regulates a growth in response to biotic and abiotic stress [11]. C₂H₄ is a highly productive source of tropospheric O₃ in urban air [12] and a precursor to SOA formation [13]. According to modern concepts, C₂H₄ is not formed in the atmosphere during chemical transformations [14]. Total atmospheric emissions of C₂H₄ are approximately 20 Tg/year, half of which come from biomass combustion; the remaining significant sources are industrial and biogenic (including oceanic) emissions [14,15]. Ethylene is widely used as a chemical feedstock, having the highest production level of all industrially produced organic compounds. In agriculture, it is an accelerator of ripening processes [11]. The main C₂H₄ removal mechanisms from the troposphere are the reaction with ozone and the hydroxyl radical (OH), and transport to the stratosphere. The lifetime of ethylene in the troposphere varies significantly with time and latitude, it ranges from a few hours in summer to about four days in winter [7,16]. It should be noted that the number of experimental studies of atmospheric ethylene is not numerous due to its low concentration in the atmosphere and, as a consequence, due to the difficulty of its detection, especially for background conditions. For the same reason, published quantitative estimates of sources and sinks are characterized by significant uncertainties [11].

The main goal of this study is to investigate the feasibility of using atmospheric monitoring data on NH₃ and C₂H₄ total columns, obtained from ground-based FTIR observations of direct solar radiation, to assess air quality in the context of its impacts on human health and the ecosystem.

The specific objectives of this paper are as follows:

-

to present the strategies that are used to retrieve the NH₃ and C₂H₄ total column (TC) in the atmosphere from high-resolution Fourier Transform Infrared (FTIR) spectra of direct solar radiation recorded at the St. Petersburg State University (SPbU) atmospheric monitoring station;

-

to analyze the uncertainty budget of the retrieved NH₃ and C₂H₄ TCs;

-

to analyze the long-term trend, annual cycle and anomalies of NH₃ and C₂H₄ using results of the long-term atmospheric FTIR monitoring of ammonia and ethylene (2009–2025) conducted in the suburbs of St. Petersburg which is the fourth-largest city in Europe with the population ~5.7 million people;

-

to compare results of NH₃ and C₂H₄ FTIR monitoring at the SPbU site with air quality standards and metrics applied for human health (national) and sensitive ecosystem state (international).

2. Materials and Methods

2.1. Description of SPbU Observational Site

The SPbU atmospheric monitoring station is a mid-latitude observational station (59.88° N, 29.83° E, 20 m asl), located at a distance of 2.5 km south of the coast of the Gulf of Finland (Baltic Sea) and of about 35 km west of the center of St. Petersburg. Due to the predominant westerly winds, most of the observations at the SPbU site are conducted outside the St. Petersburg pollution plume and mostly reflect the regional atmospheric background. The geographic location of the SPbU station is indicated in Figure 1 (global (a) and local (b) maps) by red circles.

The FTIR-observational site is located at the suburban SPbU campus, the territory of which includes the following terrain patterns: pedestrian roads—4%, motor roads and parking lots—12%, commercial buildings of the University campus—18%, and vegetation (grass, bushes, and trees)—66%. The relative distribution (in percents) of land resources in the regional scale (Leningrad Region) by land category is given in

Table 1. Relative distribution (in percentage) of land resources in the regional scale (Leningrad Region) by land category.

Land Category	Percent of Total Area
Forest	57
Agricultural lands	20
Lands of populated areas	3
Lands for industry, transport, communications, etc.	5
Lands of specially protected territories and objects	<1
Water bodies	13
Reserve lands	2

(according to information presented in [17]).

The primary urban sources of ammonia in St. Petersburg are wastewater treatment and stationary industrial facilities while in Leningrad Region ammonia emissions are driven by intensive agriculture and mineral fertilizer production. The region is home to major chemical facilities like EuroChem-Northwest (fertilizer manufacturer), which has a production capacity of 1 million tons of ammonia per year. In agriculture, manure management and nitrogen-based fertilizers from livestock operations are main fugitive sources of NH₃. However, official inventories for both St. Petersburg and Leningrad Region are available only for motor vehicles, which contributed ~2300 and ~800 tons of NH₃ in 2024, respectively [18,19].

While there is no official data on ethylene emissions, scientific literature indicates that ethylene typically accounts for less than 5% of total anthropogenic non-methane VOC emissions in urban environments [20]. In 2023 total annual emissions of volatile organic compounds (VOCs) from stationary sources and motor vehicles were 12,500 tons in Saint Petersburg [18] and 59,500 tons in Leningrad Region [19], therefore the approximate values of total annual C₂H₄ emission are less than 625 and 2975 tons respectively.

2.2. FTIR System

Since 2009, the FTIR system for atmospheric applications has been used to record spectra of direct solar radiation in the mid-IR range [21]. The SPbU FTIR system consists of:

-

the Bruker IFS125 HR high-resolution Fourier transform spectrometer (FTS) (Bruker, Billerica, MA, USA) installed in a thermostatted room. The highest spectral resolution of the Bruker IFS 125HR is Δν = 0.0019 cm⁻¹;

-

the original solar tracking system developed at the Dept. of Atmospheric Physics of St. Petersburg State University.

The routine monitoring of FTS alignment is being regularly (approximately once a month) carried out by HBr (or N₂O) cell spectra measured for optical path difference OPD = 180 cm (Δν = 0.005 cm⁻¹) using a liquid nitrogen (LN)-cooled Indium Antimonide (InSb) detector and a globar (a standard internal middle IR light source of Bruker IFS 125HR). For the acquisition of HBr cell spectra we set KBr beamsplitter in FTS, while for N₂O cell spectra, we replace it with CaF₂ beamsplitter. Retrievals of modulation efficiency (ME) and phase error (PE) characterizing the quality of FTS alignment are being performed by LINEFIT software [22]. Results of ME and PE retrievals for 2012–2025 demonstrating good alignment of FTS are given in Figure 2.

These FTIR measurements are carried out in collaboration with the Infrared Working Group (IRWG) of the Network for the Detection of Atmospheric Composition Change (NDACC). The advantage of FTIR observations is the ability to simultaneously determine up to 30 gas components of the atmosphere (https://www2.acom.ucar.edu/irwg, accessed on 10 April 2026).

FTIR monitoring of atmospheric gas composition is carried out under cloudless skies or with small cloud cover. In total, FTIR observations at the SPbU station were conducted over 1045 days from 2009 to 2025 (~70–80 days per year). Gaps in data are due to unsuitable weather conditions or failures of the FTIR system. The typical setup of FTIR system for the NH₃ and C₂H₄ monitoring using the mid-IR spectra of direct solar radiation at the SPbU station is given in

Table 2. Typical SPbU FTIR system setup for mid-IR atmospheric measurements of NH₃ and C₂H₄.

Detector	LN-Cooled MCT
Beamsplitter	KBr
Fieldstop, mm	2.0–2.5
Δν, cm−1 (OPD, cm)	0.005 (180)
Registered spectral range, cm−1	650–1400 (optical filters F3 and F3*)
Number of scans	6–10

. Currently, the spectra are recorded using the F3 optical filter in combination with the LN-cooled MCT (Mercury–Cadmiun–Telluride) detector. From 2009 to March 2016, another optical filter, F3*, with similar characteristics but providing a lower signal-to-noise ratio (SNR), was used for measurements in this spectral range.

To achieve a better SNR, the interferograms are accumulated and then averaged. The number of scans for most atmospheric FTIR measurements at the SPbU station is set as 8–10 (the duration is approximately 10–12 min). For observations conducted under variable cloud cover and/or at large solar zenith angles (SZAs), the number of scans can be set to 6.

2.3. Processing FTIR Spectra

The algorithm for solving the inverse problem of atmospheric remote sensing using FTIR observations is aimed at obtaining quantitative information on the total column (TC_GAS) and/or vertical profile of the volume mixing ratio of the target gas in the atmosphere based on measured IR spectra of direct solar radiation. The theoretical fundamentals for solving ill-posed problems of atmospheric remote sensing were developed by Tikhonov [23] and Rogers [24] in the 1960s and 1970s. Optimal estimation (OE) and Tikhonov–Phillips regularization (T-P) are the two most common approaches implemented in SFIT4 software [25], which we used to process the FTIR spectra [26].

In developing and adapting the retrieval strategies for deriving TC_NH3 and TC_C2H4 from high-resolution FTIR spectra, we relied on published works [7,8,9,10,27,28]. To ensure comparability of our data with the results of other FTIR stations of IRWG NDACC, we used the unified (homogeneous, consistent with other IRWG NDACC stations) a priori and input information when processing the spectra:

(1)

A priori vertical distributions of NH₃ and the corresponding a priori covariance matrices were formed based on simulations of the GEOS-CHEM model [10]. For C₂H₄, as well as interfering components (indicated in column 4 of

Table 3. Information on the NH₃ and C₂H₄ retrieval strategies used for processing FTIR spectra measured at the SPbU station.

Target Gas	Spectral Intervals, cm−1	Spectroscopic Linelist	Retrieved Interfering Gases	Regularization Type
NH3	929.40–931.40 962.10–970.00	ATM 20	H2O, O3, CO2, CO2636, N2O, CO2628, HNO3	OE
C2H4	948.80–952.40	ATM 20	CO2, H2O, COF2, N2O, NH3, O3, SF6	T-P

), we used the data from the global chemical-transport atmospheric model WACCM v.6 (Whole Atmosphere Community Climate Model) [29].

(2)

The NOAA/NWS/National Centers for Environmental Prediction (NCEP) data on temperature profiles and geopotential heights at 18 levels from 1000 mb to 0.4 mb were used as input information on atmospheric temperature and pressure (https://www-air.larc.nasa.gov/missions/ndacc/data.html?NCE12.00UTCP=ncep-list data access up to 10 April 2026 and https://www-air.larc.nasa.gov/missions/ndacc/data.html?NCEP_GFS=gfs-list data access after April 2025; accessed on 10 April 2026).

(3)

Spectroscopic data, whose accuracy largely determines the reliability of the retrieval results for the target gas, were taken from the HITRAN [30,31] and ATM [32] databases of various versions. Also, the data on solar (Fraunhofer) spectral lines [33] were used.

The information on the retrieval strategies used to determine TC_NH3 and TC_C2H4 is given in Table 3. The second column in Table 3 indicates the spectral ranges. The third column contains information on the spectroscopic database. The fourth column lists the interfering trace gases which were retrieved simultaneously with NH₃ and C₂H₄. And the fifth column indicates the type of regularization.

Typical examples of measured and calculated spectra, as well as the differences between them in the spectral ranges used for TC_NH3 and TC_C2H4 retrievals are shown in Figure 3a,b.

The formalism commonly used to solve the ill-posed inverse problems of atmospheric remote sensing involves calculating the so-called averaging kernel matrix [34,35]. This matrix helps to estimate the vertical resolution of the retrieved profiles and to derive some other important characteristics. In particular, the averaging kernels (AVKs) which are the rows of this matrix show how the variations in all individual elements of the true profile will contribute to the value of each element of the retrieved profile. AVKs for the mixing ratio of NH₃ and C₂H₄ are illustrated in Figure 4a,b. The AVKs are functions of altitude (see Figure 4); in an ideal scenario, the retrieval would have an AVK equal to unity in the region of interest (altitude) and zero beyond it. AVKs characterize the degree of smoothing of the true profile in the procedure of solving the inverse problem. From Figure 4, it is evident that our retrieval strategies have the highest sensitivity to the NH₃ and C₂H₄ contents in the troposphere and stratosphere (up to altitudes of ~30–40 km for NH₃ and up to ~15 km for C₂H₄) with maximum sensitivity in the middle and upper troposphere (at altitudes of ~4–10 km). It should be noted that the trace of the AVK matrix characterizes the number of the degrees of freedom for signal (DOFS), which in our case can be interpreted as the number of atmospheric layers where mixing ratios for the target gas are determined independently. The DOFS value shows whether only the TC_GAS can be determined (for DOFS < 2) or whether elements of the gas vertical distribution in the atmosphere can be retrieved (for DOFS ≥ 2). Thus, in our case, FTIR measurements can only provide the information on TCs of NH₃ and C₂H₄, since the DOFS values for both target gases are ~1.0 (see

Table 4. Average values of TC_GAS (±σ), X_GAS (±σ), RMS (±σ), DOFS (±σ), obtained as a result of processing FTIR spectra for the entire measurement period.

Target Gas (Measurement Period, Optical Filter)	TCGAS, 1015 molec/cm2	XGAS, pptv	RMS, %	DOFS
NH3 (2009–2025)	3.50 ± 3.30	163 ± 156	0.32± 0.26	1.02± 0.08
NH3 (2009–March 2016, F3*)	4.38 ± 3.25	210 ± 160	0.49± 0.27	1.01± 0.08
NH3 (April 2016–2025, F3)	2.95 ± 3.29	140 ± 150	0.21± 0.19	1.03± 0.07
C2H4 (April 2016–2025, F3)	1.27 ± 1.25	59 ± 58	0.21± 0.07	1.1± 0.1

below).

2.4. Primary Data Analysis and Error Budget

The results of the FTIR spectra processing have shown that the C₂H₄ retrieval setup applied to spectra recorded with the F3* optical filter does not allow the SFIT4 iterative procedure to converge due to low SNR value. Therefore, we processed spectra only for the period from April 2016 to 2025 when F3 filter with higher SNR was used. Table 4 presents the most important results including mean values of the total columns of NH₃ and C₂H₄, the column-averaged dry-air mole fraction (X_GAS) of NH₃ and C₂H₄, the root mean square difference between the measured and calculated spectra (RMS), and DOFS. In the case of NH₃, we separately show the results for filters F3 (April 2016–2025) and F3* (2009–March 2016). Column-averaged dry-air mole fraction of the target gas X_GAS was calculated as the ratio of the TC_GAS and the dry pressure column (TC_{dry_air}) [36]:

X_GAS = TC_GAS/TC_{dry_air}

(1)

TC_{dry_air} = (P_s/g(φ) − TC_H2O·m_H2O)/m_{dry_air}

(2)

where m_{dry_air}—the molecular mass of dry air (28.964 g mol⁻¹);

m_H2O—molecular mass of H₂O (18.02 g mol⁻¹);

P_s—surface pressure (hPa) which is taken from NCEP input data used in the retrievals;

g(φ)—the latitude-dependent gravitational acceleration;

TC_H2O—the H₂O total column obtained as a result of separate retrieval.

For convenience, we also provide here the formula for converting concentration q from ppbv to mg/m³:

q_mg/m3 = q_ppbv·P·m_GAS/(10³·R·T)

(3)

where m_GAS—the molecular mass of studied gas (g mol⁻¹);

P—pressure (Pa);

T—temperature (K);

R—ideal gas constant (8.314 J K⁻¹ mol⁻¹).

Average values of TC_GAS and X_GAS, given in Table 4, are the estimates of the regional background levels of TC_{GAS_BG} and X_{GAS_BG} for our station. Since NH₃ and C₂H₄ are most abundant in the troposphere, the X_GAS values can be considered as the background tropospheric concentrations of NH₃ and C₂H₄. High values of σ for TC_GAS and X_GAS (see Table 4) reaching ~100% for NH₃ and C₂H₄ are caused by high reactivity and deposition rates of these gases.

When assessing the uncertainty budget of the retrieved total column of target gas in the atmosphere, we followed the formalism developed by Tikhonov and Rodgers [23,24] which is incorporated into the SFIT4 retrieval tools. A detailed description of the main sources of systematic, random, and smoothing errors which we took into account in our uncertainty analysis can be found in [21]. In our case, it is assumed that the smoothing error is random and characterizes the uncertainty caused by the limited vertical resolution of the FTIR technique. The average values of relative random (δ_rand) and systematic (δ_sys) errors, as well as smoothing (δ_sm) errors of NH₃ and C₂H₄ TCs, estimated as a result of processing the FTIR spectra collected at the St. Petersburg State University atmospheric monitoring station, are presented in

Table 5. Average values of relative errors of TC_NH3 and TC_C2H4, obtained for the atmospheric monitoring station of St. Petersburg State University.

Target Gas (Measurement Period, Optical Filter)	Relative Error, %
	δrand	δsm	δsys
NH3 (2009–2025)	6.7	0.2	23
NH3 (2009–March 2016, F3*)	9.5	0.1	27
NH3 (April 2016–2025, F3)	5.1	0.3	20.3
C2H4 (April 2016–2025, F3)	26	0.8	15

. This table shows that using the F3* optical filter in FTIR measurements from 2009 to March 2016 results in higher errors of TC_NH3 compared to corresponding errors obtained for F3 filter.

The obtained values of the TC_NH3 error components presented in Table 5 are consistent with the results of independent studies. The values of random, systematic, and smoothing errors estimated from FTIR measurements at the Hefei station (China) are 8.1 ± 5.8%, 27.8 ± 19.7%, and 0.6 ± 0.4%, respectively [37]. For C₂H₄, the δ_rand and δ_sys values for the SPbU station are generally consistent (taking into account the difference in the geographical location of the stations) with the corresponding errors estimated for the high-latitude Eureka station (Canada): δ_rand = 64% and δ_sys = 17.9% [7].

2.5. Atmospheric Dispersion Modeling

In our study, modeling WF plume evolution was performed by HYSPLIT (Hybrid Single-Particle Lagrangian Integrated Trajectory) Dispersion Model [38,39] via the online NOAA READY Website Engine (https://www.ready.noaa.gov/HYSPLIT_disp.php, accessed on 10 April 2026), model setups and details for this run are given in Appendix A (Figure A1). The HYSPLIT model, which is a comprehensive computer system used to simulate the atmospheric transport, dispersion, and deposition of pollutants and hazardous materials over local to global scales. The HYSPLIT is a hybrid mathematical framework incorporating Lagrangian and Eulerian approaches. The key capabilities and features of the HYSPLIT are as follows:

-

Trajectory analysis: model computes forward air mass trajectories to determine/forecast the path of pollutant, and backward trajectories to track an air mass back in time to establish its source-receptor relationship;

-

Atmospheric dispersion modeling: the model simulates pollutant mass distribution by releasing it as point-mass particles, growing 3D cylindrical puffs, or a hybrid combination of both;

-

Advanced Physics: the model incorporates horizontal and vertical complex wind shear, vertical diffusivity profiles, chemical transformation, radioactive decay, and dry or wet deposition;

-

Meteorological flexibility: the model utilizes previously gridded, pre-processed regional or global binary meteorological data from agencies like NCEP, utilizing either historical archives or future forecast fields.

The model serves as a useful tool for environmental safety, aviation, and regulatory policy and has the following common practical applications: analysis of natural and anthropogenic hazardous release events (tracking the concentration and deposition of radionuclides from nuclear accidents, forecasting the movement of volcanic ash, tracking wildfire smoke plumes, and mapping windblown dust storms); air quality applications (tracking urban pollutants like PM2.5, stationary factory emissions, and identifying transboundary transport of toxic atmospheric components to determine air pollution levels); inverse modeling (utilizing experimental data on atmospheric composition to establish the location of unknown emission source and its emission intensity). For clarity, the methodology used in this study is illustrated in Figure 5.

3. Results Analysis and Discussion

The results of long-term FTIR measurements are shown in panels (a) of Figure 6, Figure 7 and Figure 8 for NH₃ and in panels (b) of Figure 6, Figure 7 and Figure 8 for C₂H₄. The TC_NH3 is characterized by significant spatiotemporal variations; for example, its average level at the SPbU station in 2009–2017 was (3.7 ± 3.2)·10¹⁵ molec/cm², which is 2.3 times lower than the corresponding value (8.4 ± 8.6)·10¹⁵ molec/cm² observed in Paris over the same period [9]. In Bremen, the average total ammonia level between 2004 and 2013 was even higher, at (13.75 ± 4.24)·10¹⁵ mol/cm² [8]. Comparison of the average TC_NH3 observed at SPbU site in 2017–2020 with similar results obtained at the Hefei FTIR station [37] showed a more than 6-fold difference in the results: ~2.91·10¹⁵ mol/cm² (SPbU station) and ~1.82·10¹⁶ mol/cm² (Hefei station). Such differences are explained by the lower intensity of NH₃ sources in the Northwest of the Russian Federation compared to the territories of the European Union and China.

The average value of TC_C2H4 observed at the SPbU station in 2016–2025 is (1.27 ± 1.25)·10¹⁵ mol/cm² which is in good agreement with the ranges of TC_C2H4 reported in the literature. Thus, for the background high-latitude Canadian station Eureka [7], spectroscopic measurements showed average values at the level of ~6·10¹⁴ mol/cm²; for the Pasadena station, located in the suburbs of greater Los Angeles, the obtained values are almost two orders of magnitude higher (~2·10¹⁶ mol/cm²) [28]. The TC_C2H4 levels less than 1·10¹⁵ mol/cm² are considered characteristic of an unpolluted atmosphere [28].

The results of NH₃ and C₂H₄ monitoring allowed us to analyze long-term trends, annual cycles, and anomalies of X_NH3 and X_C2H4 observed in the atmosphere of suburban territory of St. Petersburg. If initially, as a result of solving the inverse problem, we obtain TC_GAS values, then switching to the X_GAS values minimizes the influence of measurement conditions (by excluding of the atmospheric pressure influence), which is important for FTIR observations with a limited number of measurement days [40]. It is for this reason that the further analysis is mainly focused on the X_NH3 and X_C2H4 values.

3.1. Long-Term Trends

To study the long-term trends, we used the original approach proposed in [21], which included the following key steps:

-

preliminary analysis of the X_GAS time series including filtering of outliers;

-

harmonic analysis of the roughly detrended non-even X_GAS time series using the Lomb–Scargle technique followed by estimation of the optimal number of harmonics (N) using the cross-validation method. Here we assume that X_GAS time series can be approximated as a model function F_X = “linear trend + N harmonics”:

$$F_X\left(t\right)=a+b\cdot t+{\displaystyle \sum _{i=1}^N{c}_i}cos\left({\alpha }_{i}t+{\phi }_i\right)$$

(4)

where t is time;

a, b, c_i, α_i and φ_i are parameters determined by the least squares method; parameter b is an estimate of the X_GAS linear trend;

-

estimation of the X_GAS linear trend and its uncertainty using bootstrapping technique.

The results obtained using this approach showed a decrease in the X_NH3 and X_C2H4 with the rate (−2.3 ± 0.2)%/year for the seventeen-year period 2009–2025 and with the rate (−2.2 ± 0.4)%/year for the ten-year period 2016–2025, respectively. The values of the X_NH3 and X_C2H4 long-term trends are statistically significant with a confidence level of 99%. According to satellite measurements carried out by the IASI (Infrared Atmospheric Sounding Interferometer) instrument [41], the NH₃ trend for the entire territory of Russia in 2008–2018 was negative and amounted to (−4.11 ± 0.80) %/year. For comparison, the X_NH3 trend estimated using our ground-based FTIR measurements for the period 2009–2018 amounted to (−3.9 ± 1.5) %/year, which is in good agreement with the above results of the IASI monitoring. The comparison of the X_NH3 trends for 2009–2018 and 2009–2025, obtained at the SPbU station, shows that the NH₃ decline rates slowed down by approximately two times from −3.9%/year to −2.3%/year.

The number of studies devoted to long-term monitoring of C₂H₄ is small due to its low atmospheric concentration and, consequently, due to the difficulty of detecting this trace gas, especially under background and near-background conditions. Toon et al. [28] provided the following estimate of the C₂H₄ total column trend for Pasadena site: a ~3-fold decrease over 25 years (quote: “Despite the increasing population and traffic in southern California, a factor 3 decrease in ethene column density is observed over JPL over the past 25 years…”). This is approximately −2.6%/year, which is consistent with our results (−2.2 ± 0.4%/year) within the error limits.

3.2. Annual Cycle

The mean annual cycles of X_NH3 and X_C2H4 shown in Figure 9 were obtained as follows: at the first stage, the long-term trend was excluded from the corresponding time series of individual measurements, then the daily and monthly means were calculated, and finally monthly means for each month were averaged over all years of FTIR measurements. Averaging periods for X_NH3 and for X_C2H4 were 2009–2025, and 2016–2025, respectively. Let us characterize the obtained mean annual cycles of X_NH3 and X_C2H4 shown in Figure 9:

(1)

The average amplitude of the X_NH3 annual cycle is ~95 pptv; the maximum of X_NH3 is observed in the warm season with a peak ~207 pptv in May, the minimum—in winter (~16 pptv in December). On average, the relative monthly variability of X_NH3 is 30–70% and the most significant variations in X_NH3 were observed in March and October. This nature of the X_NH3 annual cycle is due to the influence of spring-summer emissions from agricultural production in the nearby region, since livestock and crop production are responsible for ~70–80% of the total ammonia input into the atmosphere, as well as emissions from wildfires [3].

(2)

The average amplitude of the X_C2H4 annual cycle is ~40 pptv; the peak of X_C2H4 occurs in January (~116 pptv), while the minimum can be observed in different months of the warm season. In our case, the minimum was usually registered in May (~38 pptv). The highest variability of X_C2H4 was observed in June. Such ethylene variability throughout the year is caused by the seasonal dependence of the main mechanism for removing C₂H₄ from the troposphere, which is the reaction of C₂H₄ with the hydroxyl radical OH.

Seasonal variations in NH₃ and C₂H₄ concentrations for different geographic locations can vary significantly depending on specific sources and sinks of the target gases [9,28]. For example, for Paris [9], the mean annual cycle of TC_NH3 based on FTIR measurements in 2009–2017 has two maxima in March and August; the lowest values are detected in winter with a minimum in January, which is generally consistent with the results obtained for the SPbU site. For ethylene, according to ACE-FTS satellite measurements [15], average concentrations in the free troposphere are ~50 pptv. Seasonal variations in C₂H₄ for the Northern Hemisphere are characterized by a maximum (~100–200 pptv) in winter—early spring and a summer minimum (~30 pptv in the absence of powerful wildfires), which is in good agreement with the results of our ground-based FTIR observations.

3.3. Anomalies Analysis

The long-term time series of NH₃ and C₂H₄ presented in Figure 7 include the dates when anomalously high levels of X_NH3 and X_C2H4 were observed. When selecting specific cases for further analysis, we used the following criteria:

-

the X_GAS value must exceed the “monthly mean X_GAS value + 3σ” level, where σ is the standard deviation (for the corresponding monthly period);

-

during the measurement day, anomalously high X_GAS values must be observed at least twice (we did not consider isolated anomaly of X_GAS as it may be an outlier), or high X_GAS values must be observed on nearby dates.

Table 6. Characteristics of the NH₃ and C₂H₄ anomalies registered at the SPbU site: the dates; PBL height averaged during the period of FTIR observations: the corresponding peak values of TC_GAS and X_GAS, and the calculated concentrations of NH₃ and C₂H₄ in the boundary layer (q_{BL_NH3} and q_{BL_C2H4}).

Date	PBL Height, m	NH3			C2H4
		TCNH3_MAX, 1015 molec./cm2	XNH3_MAX, pptv	qBL_NH3, ppbv	TCC2H4_MAX, 1015 molec./cm2	XC2H4, pptv	qBL_C2H4, ppbv
13 May 2010	168	14.0	810	31	-	-	-
14 May 2010	167	17.2	807	31	-	-	-
30 June 2010	137	12.1	563	24	-	-	-
8 July 2010	81	14.9	701	55	-	-	-
19 May 2013	170	22.7	1064	43	-	-	-
10 April 2014	1017	17.4	804	6	-	-	-
27 July 2017	507	-	-	-	2.89	137	1.2
25 September 2017	202	-	-	-	4.95	226	7.9
17 October 2017	881	18.1	851	10	26.4	1243	12
22 May 2018	302	18.3	849	18	-	-	-
11 July 2018	574	-	-	-	3.51	164	1.5
10 August 2018	901	13.1	609	4	-	-	-
28 August 2018	486	-	-	-	9.48	442	7.1
19 September 2018	681	21.8	1030	11	-	-	-
26 April 2019	251	17.9	833	23	-	-	-
22 July 2019	587	-	-	-	2.79	130	1.0
5 July 2022	657	-	-	-	2.58	121	0.7
20 February 2023	257	-	-	-	5.14	244	5.4
24 April 2023	413	24.8	1169	21	-	-	-
21 August 2023	763	20.2	946	9	-	-	-
3 July 2025	476	11.9	562	7	-	-	-

shows the results of this selection: the dates when anomalies of X_NH3 and X_C2H4 were detected and the corresponding peak values of X_NH3 and X_C2H4 on these days (highlighted by red diamonds in Figure 7). The observed positive anomalies in X_NH3 and X_C2H4 are caused by atmospheric emissions from significant natural and anthropogenic pollution sources. At our station, the influence of both types of sources is present, and during certain periods, it can be significant.

It should be noted that for both gases, the DOFS value is ≈1. This means that when processing spectra measured in the pollution plume, the inverse problem solving algorithm can determine the total number of gas molecules (which absorb solar IR radiation) along the solar ray trace, but cannot determine the vertical location of atmospheric layers with anomalously high concentrations of NH₃ and C₂H₄. As already mentioned, the majority of NH₃ and C₂H₄ are contained in the troposphere, and most FTIR measurements are conducted under regional background conditions. Therefore, the average values of X_NH3 and X_C2H4 characterize the mean tropospheric concentrations of these gases. Since NH₃ and C₂H₄ have a short atmospheric lifetime, their most significant anomalies recorded at our station are due to the emissions of powerful nearby ground-based sources (these gases are not formed in the atmosphere). In this case, the bulk of the pollutants will be concentrated in the planetary boundary layer (PBL), where the concentration of NH₃ and C₂H₄ can be estimated using the difference between the recorded anomalies of TC_NH3 and TC_C2H4 and their monthly mean levels (considered as a background levels). Using the calculated background monthly mean tropospheric levels TC_{GAS_BG} and X_{GAS_BG}, and taking ERA5 hourly data on H_PBL [42] for our region (see Table 6, second column), we can approximately estimate NH₃ and C₂H₄ mean concentrations in PBL for the registered anomalies given in Table 6. For this, we derive the following formula:

q_{BL_GAS} = X_{GAS_BG} + (TC_{GAS_MAX} − TC_{GAS_BG})·k·T/(H_PBL·p)

(5)

where X_{GAS_BG}—background monthly mean tropospheric value of X_{NH3_BG} and X_{C2H4_BG};

TC_{GAS_MAX}—peak values of TC_NH3 and TC_C2H4 (see Table 6) (molec./m²);

TC_{GAS_BG}—background monthly mean tropospheric value of TC_{NH3_BG} and TC_{C2H4_BG} (molec./m²);

k—the Boltzmann constant (1.38·10⁻²³ J/K)

T—temperature at ground level (K);

H_PBL—mean PBL height during FTIR observations according to ERA5 data (see Table 6, second column);

p—atmospheric pressure at ground level (Pa).

The results of calculations of q_{BL_NH3} and q_{BL_C2H4} are given in the fifth and eighth columns of Table 6.

Over the entire measurement period at SPbU station, the NH₃ and C₂H₄ anomalies were observed for 21 and 8 days, respectively. High concentrations of both gases were recorded simultaneously only once, on 17 October 2017. Based on the information presented in Table 6, we note the following features of the distribution of X_NH3 and X_C2H4 anomalies throughout the year:

-

Only one event was recorded during the cold season (November–March), this case is exclusively X_C2H4 anomalies;

-

A total of 20 events, 14 of which were X_NH3 anomalies, were recorded during the warm season from April to October;

-

About 30% (7 events) of the total number of NH₃ anomalies were detected in April–May.

The latter is explained by the observed spring peak in NH₃ emissions from agricultural activities (fertilizer application and livestock emissions), as well as the peak of spring wildfires. Since both phytotoxicants are intensely emitted during biomass burning and their local concentrations in pollution plumes can simultaneously reach high levels, this can negatively impact plant growth, especially in the spring. Although our FTIR measurements during the spring did not record simultaneous anomalies in X_NH3 and X_C2H4, such events can still be observed under certain conditions. For example, on 17 October 2017 we recorded the passage of a pollution plume from forest fires over the SPbU station. The highest TC_C2H4 value for the entire period of measurements at our station was recorded on this day (see Table 6). To confirm that the NH₃ and C₂H₄ anomalies on 17 October 2017 were associated specifically with the passage of the wildfire pollution plume, we used the data on the CO, HCN, and C₂H₆ TCs (biomass burning products) in the atmosphere, obtained from FTIR measurements at our station in 2009–2025 [27,43]. Special attention was paid to changes in X_HCN (for the fall of 2017), since high concentration of HCN is the most characteristic indicator of the biomass burning products presence in the air [27,43,44,45]. Figure 10 shows the daily average X_GAS values of NH₃, C₂H₄, and HCN recorded over the period of approximately 20 days before and 20 days after 17 October 2017. It is clearly seen that on 17 October 2017, there was a synchronous significant increase in X_GAS concentrations for all three atmospheric gases: by a factor of ~5 for NH₃, by a factor of ~20 for C₂H₄, and by a factor of ~6 for HCN.

In the late summer and fall of 2017, powerful forest fires were observed in the high latitudes of the Northern Hemisphere. This phenomenon is discussed, for example, in the article by Wizenberg et al. [7], who analyzed the influence of North American forest fires on the gaseous composition of the atmosphere over the Canadian FTIR station Eureka (80.05° N, 86.42° W). Information on the forest fire location from EFFIS database [46] for 12–17 October 2017 showed that during this period, a number of wildfires were also registered in the territory of Fennoscandia (see Figure 11):

-

In Sweden, 14 fires were recorded, including three of them near 58.35° N, 12.37° E and four of them near 60.13° N, 16.17° E;

-

In Finland, 10 fires were recorded, localized in the areas of 60.29° N, 25.52° E and 64.64° N, 24.42° E;

-

In Norway, 2 fires were recorded: at 60.23° N, 10.35° E and 59.12° N, 9.61° E, with the peak of fires occurring on 16–17 October.

The power of the most intense forest fires in Finland on 17 October 2017 was estimated at 128 MW (60.29° N 25.52° E), 103 MW (60.30° N 25.52° E) and 88 MW (60.29° N 25.52° E). Simulation of atmospheric dispersion using the HYSPLIT model [38,39] made it possible to establish that the reason of the X_NH3, X_C2H4 and X_HCN anomalies, observed on 17 October 2017 at the SPbU station, was the forest fires in the Helsinki area (located approximately 250 km to the north-west from the observational site; marked in Figure 11 with a purple sign). Figure 12 shows the location of the pollution plume in the atmosphere above the SPbU station (indicated by a red dot) during the period from 9 to 11 UTC on 17 October 2017, when we recorded peak values of TC_NH3 and TC_C2H4. The results of modeling show that during this time, FTIR measurements were conducted in the central part of the plume with the highest biomass-burning product concentrations (see Figure 12).

3.4. Air Quality Metrics in Comparison with the Results of NH₃ and C₂H₄ FTIR Monitoring

If for C₂H₄ the maximum values of X_{C2H4_MAX} = 1243 pptv and q_{BL_C2H4} ≃ 12 ppbv are obtained on the same date (17 October 2017), then for NH₃ the maxima of X_{NH3_MAX} = 1169 pptv and q_{BL_NH3} ≃ 55 ppbv were recorded on different measurement days (24 April 2023 and 8 July 2010, respectively). For ammonia and ethylene, classified in the Russian Federation as class 4 and 3 hazard substances, the given values do not exceed the maximum permissible concentration (MPC) levels established in the Russian Federation for the atmospheric air of urban and rural settlements, which are the following [47]:

(1)

For the NH₃ MPC, the maximum one-time value (exposure up to 30 min) is 0.2 mg/m³ (287 ppbv), and the daily average value (exposure for 24 h) is 0.1 mg/m³ (143 ppbv);

(2)

For the C₂H₄ MPC, the maximum one-time value (exposure up to 30 min) is 3.0 mg/m³ (2.6 ppmv).

These values are established taking into account the toxic effects of NH₃ and C₂H₄ on human health. Ethylene, when exceeding the MPC, can have a narcotic effect, causing headaches, dizziness, suffocation, and circulatory problems. Ammonia is a substance with a reflex-resorptive effect, causing severe irritation of the upper respiratory tract. At the same time, the presence of NH₃ in the atmosphere also has a protective effect on human health against population exposures to acidic aerosols. This is particularly important in polluted atmospheres, where if NH₃ levels were controlled, the negative health impacts of air pollutants from fossil fuel combustion would be exacerbated.

In other countries, the established standards for NH₃ and C₂H₄ may differ from those in Russia. For example, in European countries, the critical concentration level of NH₃ (CLE_NH3), developed for sensitive ecosystems, is also used [3,48]. CLE_NH3 is the concentration of atmospheric ammonia above which, according to current understanding, direct adverse effects on an ecosystem may occur [48,49,50]. These effects include impacts on biodiversity (e.g., on species composition and community composition, such as a decrease in the frequency or number of nitrogen-sensitive species or an increase in the number of nitrogen-tolerant species) and impacts on physiological indicators (e.g., an increase in nitrogen content or nitrogen leaching from the soil). Establishing of CLE_NH3 in international studies is more focused on protecting ecosystems and biodiversity, rather than solely on human health. Russian regulations, however, emphasize sanitary and hygienic aspects—protecting public health. As a result, although MPCs in Russia set maximum concentration limits for substances in the air similar to international standards—they are not fully equivalent to CLEs in terms of their objectives and determination methodology. It should be noted that a promising direction for future research based on the obtained results of atmospheric FTIR monitoring is the most detailed study of quantitative indicators of environmental impact and environmental metrics, including the use of additional experimental and model data.

Although Critical Level (CLE) criteria are not in use in Russia, we will compare the results of our FTIR NH₃ measurements with established CLE_NH3 levels for European countries. It should be noted that in 2009–2010, stricter European standards were introduced for NH₃: CLE_NH3 = 1–3 μg/m³ (1.4–4.3 ppbv), which is significantly lower than the CLE_NH3 = 8 μg/m³ (11 ppbv) previously established in the 1980s [3,48,51]. The results that we obtained for the mid-tropospheric concentration (X_{NH3_BG} ≃ 160 pptv) are typical for the conditions of the suburbs of St. Petersburg and are more than an order of magnitude lower than CLE_NH3, but the situation is changing for observations carried out in the pollution plumes of intense emission sources, when q_{BL_NH3} can reach 55 ppbv. Since the lifetime of NH₃ is less than 24 h, the anomalies which we recorded demonstrated the presence of both permanent (agricultural enterprises, landfills, etc.) and irregular (wildfires) significant regional sources of ammonia. It should be noted that the number of detected NH₃ anomalies significantly exceeds the number of C₂H₄ anomalies. Thus, despite the generally low levels of NH₃ pollution in the suburb of St. Petersburg, NH₃ concentrations in close vicinities to existing regional sources can permanently exceed CLE_NH3, affecting the state of sensitive ecosystems.

Ethylene is a key plant hormone produced by plants in response to abiotic stress. It is crucial for plant growth and development under various abiotic stresses, including salinity, hypoxia, and heat. C₂H₄, depending on its concentration in the air, can have both positive and negative effects on plants. Due to its importance in agriculture, the biochemistry of ethylene has been well studied by plant physiologists [52]. Ethylene concentrations above 0.01–0.05 ppmv can inhibit flowering, shorten internode length, increase branching, initiate fruit ripening, cause leaf and flower senescence and abscission, chlorosis (yellowing) of leaves, and promote the formation of adventitious roots. Crop sensitivity to C₂H₄ varies, but there are extremely sensitive plant species that react sharply to a short-term (8–24 h) increase in concentration to 0.01 ppmv (e.g., tomatoes and Cuphea hyssopifolia) [52].

The mid-tropospheric concentration X_{C2H4_BG} for the suburbs of St. Petersburg is ≃60 ppbv, which is approximately three orders of magnitude lower than the specified thresholds of 0.01–0.05 ppmv. Unlike NH₃, however, a relatively short exposure time to C₂H₄ concentrations at the level of 0.01 ppmv is sufficient to negatively impact certain plant species. The observed plume of wildfires on 17 October 2017 with mean concentrations in PBL of q_{BL_C2H4} ≃ 12 ppbv could have such an impact. It should be noted that, since the main growing season for open-ground plants in the northwestern region of the Russian Federation usually ends by this time, a similar impact may also affect greenhouse crops. Over the ten-year period of 2016–2025, eight cases with extremely high levels of X_C2H4 were registered at the St. Petersburg State University station, only two of them were observed in cold season (17 October 2017 and 20 February 2023) and six events were detected in the summer (the growing season). The levels of X_C2H4 observed in the atmosphere of the suburbs of St. Petersburg are low and are close to the background levels. Nevertheless, wildfires, which can lead to significant increases in the C₂H₄ concentrations in the lower troposphere during summer, appear to be an important factor that can negatively impact vegetation on a regional scale. It should be noted that the wildfire plumes usually contain several gaseous products of biomass burning, which are phytotoxicants (for example, C₂H₄, NH₃, SO₂, PAN, etc.), so in this case the inhibitory effect on plants can be mutually reinforcing.

4. Conclusions

This work demonstrates the applicability of atmospheric monitoring results of NH₃ and C₂H₄ total columns, obtained from ground-based FTIR observations of direct solar radiation, to assess air quality in the context of its impacts on human health and the ecosystems in urbanized areas. As a result of FTIR measurements, the annual cycle, long-term trends, and positive anomalies of the integral content of NH₃ and C₂H₄ in the atmosphere in a suburb of St. Petersburg agglomeration have been obtained for the period 2009–2025. These parameters have been compared with the maximum permissible concentrations (MPCs) established in the Russian Federation. The potential impact of high pollution levels of NH₃ and C₂H₄ on the region’s sensitive ecosystems has been analyzed. Our results have demonstrated that the retrievals of the NH₃ and C₂H₄ total columns in the atmosphere can be successfully accomplished by processing of FTIR spectra of direct solar radiation in three spectral ranges: 929.40–931.40 cm⁻¹ (NH₃), 962.10–970.00 cm⁻¹ (NH₃), and 948.80–952.40 cm⁻¹ (C₂H₄). The random/systematic errors of the NH₃ and C₂H₄ total column retrievals have been estimated as 6.7%/23% and 26%/15%, respectively.

Results of long-term FTIR-monitoring at the SPbU station showed that, despite the proximity of the measurement site to Russia’s second-most populous city (~5.7 million people), the observed levels of X_NH3 and X_C2H4 can be considered as being close to background levels. Thus, the average long-term levels for the SPbU station are ~163 pptv for NH₃ (2009–2025) and ~59 pptv for C₂H₄ (2016–2025), which are significantly lower than the MPCs established in Russia.

The analysis of X_NH3 and X_C2H4 time series allowed us to evaluate long-term trends, seasonal variability, and identify anomalies in X_NH3 and X_C2H4. It was shown that for both target gases, a statistically significant decrease in X_NH3 and X_C2H4 values was observed, amounting to (−2.3 ± 0.2)%/year for the 2009–2025 period and with the rate (−2.2 ± 0.4)%/year for the 2016–2025 period, respectively. The annual cycle of X_NH3 with the amplitude of ~95 pptv has a maximum in the warm season with the peak (~207 pptv) in May; the minimum is observed in winter (~16 pptv in December). For C₂H₄, the average amplitude of seasonal fluctuations is ~40 pptv, the peak of X_C2H4 occurs in January (~116 pptv), and the minimum is usually registered in May and is ~38 pptv.

Over the entire period of FTIR measurements at the St. Petersburg State University station, extremely high X_NH3 and X_C2H4 values were recorded for 21 and 8 days, respectively. During the warm season from April to October, anomalous X_NH3 and X_C2H4 values were recorded 20 times, 14 of which were X_NH3 anomalies. About 30% (seven events) of the total NH₃ anomalies occurred in April and May. For St. Petersburg and the nearby region, this may negatively impact vegetation, reducing plant resistance to frost, drought, and other stressful situations, as well as reducing CO₂ absorption during photosynthesis. Periodically recorded X_NH3 anomalies indicate the presence of intensive emission sources in the region, subjecting ecosystems in adjacent areas to constant exposure to NH₃ concentrations exceeding the critical level (CLE_NH3).

Anomalously high concentrations (X_NH3 and X_C2H4) of the target phytotoxicants were recorded simultaneously only once—on 17 October 2017. Using data on HCN total column (as a forest fire indicator) and the results of atmospheric dispersion modeling, it was shown that this pollution event was caused by the influence of biomass burning products emitted from wildfires located approximately 250 km to the north-west from the observational site in the Helsinki area (Finland). Wildfires, which can lead to significant increases in C₂H₄ concentrations in the lower troposphere (PBL), are a factor that can negatively affect vegetation (including greenhouse plants) on a regional scale. This is especially important in light of the observed climate changes, which lead to an increase in the number and area of wildfires.