Cloud Condensation Nuclei (CCN) and Ice Nucleating Particles (INP) Conversion Factors Based on Thessaloniki and Leipzig AERONET Stations Using CALIPSO Aerosol Typing ^†

Abstract

An analysis was conducted using AERONET Inversion Data at Thessaloniki and Leipzig stations. Aerosol type plays a vital role in determining their ability to act as CCN or INP, as properties such as chemical composition, morphology, and particle size influence their hygroscopic and ice-nucleating behavior. The CALIPSO mission provides global aerosol classification with vertical resolution by using backscatter intensity and depolarization ratio measurements. Aerosol typing from CALIPSO overpasses within 100 km of each selected AERONET station was used. Only pure aerosol cases (dust, polluted continental, smoke) were selected. This study combines AERONET-derived microphysical properties with CALIPSO aerosol classification to estimate particle number concentrations relevant for CCN and INP formation. The aim is to derive improved conversion factors for each aerosol type, enabling their application in future CCN and INP concentration profiles.

1. Introduction

Ground-based polarization lidar provides highly effective vertical profiles of aerosol parameters, enabling the estimation of cloud condensation nuclei (CCN) and ice-nucleating particles (INP) number concentrations, as demonstrated in multiple studies [1,2,3]. The results are based on reliable conversion factors between aerosol optical thickness and column-integrated particle size distribution based on Aerosol Robotic Network (AERONET) photometer observations. A key aspect influencing the ability of aerosol particles to function as CCN or INP is their specific aerosol type. According to Mamouri and Ansmann (2016) [2], atmospheric aerosol particles can be categorized into dust-dominated and pollution-dominated types, based on the Ångström exponent. Low Ångström exponent values (<0.5) typically correspond to larger, coarse-mode particles like desert dust, while high values (>1.6) are associated with smaller, fine-mode particles typical of urban pollution. However, dust might also be present in the fine mode [4]. A Fine Mode Fraction (FMF) threshold of less than 0.4 was employed to identify mineral dust [5]. Annual mean MODIS FMF values for mineral dust have been reported as 0.45 ± 0.05 [6]. Shin et al. (2019) [7] found that coarse-mode AOD from AERONET tends to overestimate mineral dust compared to dust AOD retrieved with PLD. Consequently, using FMF, fine-mode AOD, or Ångström exponent as proxies for identifying non-dust aerosols would systematically underestimate the contribution of non-dust particles to the total aerosol plume. In this study, aerosol type is given by Cloud-Aerosol LIDAR and Infrared Pathfinder Satellite Observations (CALIPSOs) satellite observations, based on PLDR measurements rather than FMF or the Ångström exponent. Microphysical properties are then derived from AERONET ground-based measurements. These datasets are combined to estimate particle number concentrations relevant to CCN and INP formation [8,9].

2. Materials and Methods

2.1. AERONET

AERONET is an extensive ground-based network of sun photometers (CIMEL instruments) that provides accurate measurements of aerosol optical and microphysical properties [10]. Observations are performed at multiple wavelengths, spanning from 340 to 1020 nm, using both direct solar irradiance and diffuse sky radiance. AERONET also supplies measurements of sky radiance as a function of scattering angle at four wavelengths (440, 675, 870, and 1020 nm). These data enable the retrieval of a wide range of aerosol properties, including aerosol optical depth (AOD), single scattering albedo (SSA), and particle size distribution. AERONET data are widely used as a benchmark for validating satellite-derived aerosol products and atmospheric model outputs. In this study, AERONET Version 3 Inversion Data (Level 1.5), which are subject to cloud screening, were employed. These data are publicly accessible via the AERONET portal (https://aeronet.gsfc.nasa.gov).

2.2. CALIPSO

Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP), the primary instrument aboard NASA’s CALIPSO satellite, captured global atmospheric profiles from June 2006 until its deactivation in August 2023. It used a standard dual-wavelength backscatter lidar operating at 532 and 1064 nm, a polarization channel at 532 nm [11], to provide high-resolution profiles (30 m vertical, 1/3 km horizontal) of attenuated and polarized backscatter from aerosols and clouds. Data were distributed as Level 1 and Level 2 products, with Level 2 providing aerosol classification (e.g., dust, polluted continental, smoke) [12,13] and aerosol optical properties such as optical depth, depolarization ratio, and lidar ratio. In this study, the Level 2 Aerosol Layer Product, Version 4, was utilized.

2.3. Methodology

An analysis was conducted using AERONET Inversion Data at two selected stations in Europe: Thessaloniki (40.63° N, 22.96° E), and Leipzig (51.35° N, 12.44° E), covering the period 2006–2021. To obtain a better classification of aerosols, aerosol typing from CALIPSO overpasses within 100 km of each selected AERONET station was utilized. Only cases classified as pure aerosols—namely dust, polluted continental, and smoke—were selected (Figure 1). The Aerosol Optical Depth (AOD) at 440 nm and the Ångström exponent (AE) 440–870 were used to calculate the AOD at 532 nm. Particle volume size distribution was based on 22 logarithmically spaced radius bins, with particle number concentration (n) calculated using volume concentration, particle volume, and a spectral integral width of 0.2716 [2]. Column-integrated values were defined as follows: n₆₀ (r > 57 nm, bins 2–22), n₁₀₀ (r > 98 nm, bins 4–22), n₂₅₀ (INP-relevant, bins 8–22 plus mean of bins 7 and 8), n₂₉₀ (bins 8–22), and n₅₀ (CCN-relevant, bins 1–11) [2,3]. To obtain the particle extinction coefficient (σ) and n₆₀, the AOD at 532 nm and the column n were divided by 1000 m.

Potential uncertainties in AERONET inversion products, such as retrieval assumptions and calibration drift, as well as in CALIPSO aerosol type classification due to algorithm limitations and misclassification, may influence the accuracy of CCN and INP conversion factors. To ensure a strong aerosol signal for reliable retrievals, only cases with AOD at 440 nm greater than 0.05 were included. In addition, to minimize the influence of mixed or uncertain aerosol conditions that can complicate interpretation and introduce classification errors, only pure aerosol types in the total column were selected. However, uncertainties inherent in the parameterization of the conversion factors and their potential impacts on the retrievals are described in [2].

Figure 1. CALIPSO Overpasses and Aerosol Detection (a), and Pure Aerosol Subtype (b) for Thessaloniki and Leipzig AERONET stations.

Figure 1. CALIPSO Overpasses and Aerosol Detection (a), and Pure Aerosol Subtype (b) for Thessaloniki and Leipzig AERONET stations.

3. Results

3.1. Pure Dust Aerosols

A total of 30 cases of pure dust aerosols were identified using CALIPSO data across the two AERONET stations, with 25 cases corresponding to Thessaloniki and 5 to Leipzig. For dust particles, n₁₀₀ (reservoir of CCN) and n₂₅₀ (reservoir of INP) were used. The conversion factor c₁₀₀ was found to be 13.6 ± 9.2 ${\mathrm{c}\mathrm{m}}^{-3}$, and the extinction exponent x_d was 0.86 ± 0.13, while the conversion factor c₂₅₀ was 0.256 ± 0.026 ${\mathrm{M}\mathrm{m} \mathrm{c}\mathrm{m}}^{-3}$ (Figure 2).

3.2. Pure Polluted Continental Aerosols

Based on CALIPSO observations, 77 cases of pure polluted continental aerosols were identified across the two AERONET stations—22 at Thessaloniki and 45 at Leipzig. For this aerosol type, n₆₀ (reservoir of CCN) and n₂₉₀ (reservoir of INP) were used. The conversion factor c₆₀ was found to be 43.4 ± 28.6 ${\mathrm{c}\mathrm{m}}^{-3}$, and the extinction exponent x_c was 0.78 ± 0.13, while the conversion factor c₂₉₀ was 0.149 ± 0.023 ${\mathrm{M}\mathrm{m} \mathrm{c}\mathrm{m}}^{-3}$ (Figure 3).

3.3. Pure Smoke Aerosols

CALIPSO revealed 13 instances of pure smoke aerosols. CCN-relevant column $n_{50}$ and INP-relevant column n₂₅₀ were derived from radius classes 1–11. The conversion factor c₅₀ was found to be 28.2 ± 28.7 ${\mathrm{c}\mathrm{m}}^{-3}$, and the extinction exponent x_c was 0.90 ± 0.19, while the conversion factor c₂₅₀ was 0.010 ± 0.004 ${\mathrm{M}\mathrm{m} \mathrm{c}\mathrm{m}}^{-3}$ (Figure 4). The conversion factors for each aerosol type, as determined in this study, are presented in

Table 1. Conversion factors for converting particle extinction coefficients to particle number concentrations at 532 nm, obtained from the regression analysis.

Desert dust	No. of Data	C100 (cm−3)	xd	C250 (Mm cm−3)
Thessaloniki and Leipzig	30	13.6 ± 9.2	0.86 ± 0.13	0.256 ± 0.026
Polluted continental	No. of Data	C60 (cm−3)	xc	C290 (Mm cm−3)
Thessaloniki and Leipzig	67	43.4 ± 28.6	0.78 ± 0.13	0.149 ± 0.023
Smoke	No. of Data	C50 (cm−3)	xs	C250 (Mm cm−3)
Thessaloniki and Leipzig	13	28.2 ± 28.7	0.90 ± 0.19	0.010 ± 0.004

.

4. Conclusions

This study demonstrates the effectiveness of integrating satellite-based aerosol classification from CALIPSO with long-term microphysical observations from AERONET for estimating CCN and INP concentrations. By focusing on pure aerosol types—specifically dust, polluted continental, and smoke—empirical conversion factors were derived to relate extinction at 532 nm to size-resolved particle number concentrations. In the case of mixed aerosol types, separation of the types will be performed based on the particle linear depolarization ratio, following the POLIPHON method [1,2,9], and each will then be treated as a pure type. For studies in different geographical and climatic regions, appropriate AERONET stations should be selected.

AERONET inversion data provided n₆₀, n₁₀₀, n₂₅₀, and n₂₉₀ values, which were matched with CALIPSO aerosol types to assess particle–cloud interaction potential. The conversion factors for dust aerosols are quite close to those found for Leipzig based in [2], which were c₁₀₀ = 13.9 ± 8.6 cm⁻³, x_d = 0.73 ± 0.09 and c₂₅₀ = 0.20 ± 0.03 cm⁻³. For polluted continental aerosols, the literature references for Leipzig report values c₆₀ = 25.3 ± 3.3 cm⁻³, x_c = 0.94 ± 0.03 and c₂₉₀ = 0.10 ± 0.04 cm⁻³, highlighting how these factors can vary significantly by region. According to [3], the conversion factors near fire were c₅₀ = 100 ± 50 cm⁻³, x_s = 0.75 ± 0.08 and c₂₅₀ = 0.18 ± 0.09 cm⁻³, while farther from fire sources, values were c₅₀ = 17 ± 5 cm⁻³, x_s = 0.79 ± 0.08 and c₂₅₀ = 0.35 ± 0.08 cm⁻³. These latter values are closer to those obtained for Thessaloniki and Leipzig. Distinct differences in conversion factors across aerosol types were evident: dust showed lower CCN values (c₁₀₀ = 13.6 ± 9.2 cm⁻³), and higher INP values (c₂₅₀ = 0.256 ± 0.026 cm⁻³), while polluted continental aerosols had the highest CCN-relevant values (43.4 ± 28.6 cm⁻³), consistent with known particle composition characteristics. Conversion factors for smoke particles depend on the distance from the fire source.

This framework enables global CCN and INP estimation from satellite-derived data, grounded in validated relationships from ground-based networks. Future research will aim to expand the aerosol dataset, enhance the quantification of associated uncertainties, and validate the findings through in situ measurements from both airborne and ground-based platforms.