Cosmic Rays and Aerosols in Earth’s Atmosphere and Their Role in Cloud Formation

Galactic cosmic rays (GCRs) consist primarily of photons, heavy ions, and electrons that have a wide energy range and are accelerated by the shocks, supernovas and explosions of massive stars. Notably, GCRs reach the Earth’s atmosphere after passing through interplanetary space. The amount of GCRs entering the Earth’s atmosphere depends on the strength of the solar magnetic field shielding created by solar activity [Muscheler 2013]. When solar activity is high, less amount of GCRs can reach the Earth’s atmosphere. This study examines the relationship between cosmic rays and cloud formation during the 11- year solar cycle and Forbush decreases (FDs) to understand the cause-effect connections between clouds, aerosols, and the ionisation by GCRs in the atmosphere. The 11-year solar cycle and FDs are long and short-term solar activity events that vary the amount of GCRs entering the atmosphere decreases by 3 - 20%. The 11-year solar cycle makes the change and recovers the previous state gradually over 11 years, on the other hand, FDs take over several following days [Kilifarska et al. 2020]. Other studies indicate that the 11-year solar cycle and FDs impact many atmospheric factors such as cloud fraction, aerosol optical depth (AOD), cloud condensation nuclei (CCN), water content, and cloud effective radius (r_eff ) [H. Svensmark et al. 2009; Pudovkin et al. 1995; Rohs et al. 2010; Dragić et al. 2011; J.Svensmark et al. 2016; H. Svensmark, J. Svensmark, et al. 2021]. This thesis summarizes the results of satellite and ground observation data analysis for both the 11-year solar cycle and FDs and simulations from the cloud processing model.

The Monte Carlo analysis was used for the satellite observation of atmospheric parameters data analysis during FDs. A newly calibrated satellite data in the Pathfinder Atmospheres Extended (PATMOS-x) from 1978 - 2018 indicated that cloud fraction portrayed statistically significant signal following FDs, at an achieved significance level (ASL) of 0.33%. The cloud emissivity also portrayed a highly-significant response; however, this could not be determined as physically caused by FDs, as the response was recorded beginning a week before the FDs occurred. In contrast, the cloud optical depth, integrated total cloud water over the whole column, and r_eff did not portray any significant signals in the frameworks of the applied methods. The top-of-atmosphere brightness temperature (TABT) at the nominal wavelengths of 3.75, 11.0, and 12.0 μm and surface brightness temperature (BT) portrayed significant signals. The estimated BT change from a radiative
transfer model (Fu-Liou model) portrayed consistent results with the observed changes in the cloud parameters observed during the FD events.

The moderate resolution imaging spectroradiometer (MODIS) data from 2000 to 2019 also indicated that the AOD at 550 nm over the ocean, combined dark target and deep blue AOD at 550 nm over the ocean, AOD at seven bands (470, 550, 660, 860, 1240, 1630, and 2130 nm), and AOD for nine model indexes (small 1 -4, large sea salt 5 -7, large mineral dust 8 and 9) had significant signals in the region between 30◦N and 30◦S by the Monte Carlo analysis. However, the AOD for fine mode and the number of CCN in the column at 550 did not portray enough significant signals.

Regarding the long-term changes in GCRs caused by the 11-year solar cycle, the apparent transmission record from 1958-2022 acquired at the Mauna Loa Observatory (MLO), Hawaii indicated corresponding changes with the GCRs observations recorded at the Moscow neutron monitor. When solar activity was high, transmission also increased. This agrees with other results that indicate that lesser amounts of GCRs cause less particle flux in the atmosphere.

Notably, the AOD observation acquired by MODIS indicated significant responses to the change in the amount of GCR during FDs; the aerosols measured by MODIS were larger than fine mode particles and Aitken sizes (1 -100 nm that are), more being affected by the ionisation in the atmosphere, resulting in their growth to CCN. In this study, we also investigated a possible link between ion-induced nucleation and condensation by GCRs that can increase the number of small particles and ions and also promote to grow aerosols to size up to diameter with 20-30 nm and large aerosol change which can influence AOD observations during events where the ionisation changes (FDs and 11-year solar cycle). The suggested solution is cloud processing since the cloud effectively grows aerosols. This might be able to explain how small aerosols (20-30 nm) grow to larger particles having a diameter of 300 nm appear which is mostly detectable by these satellites.

We estimated the AOD and transmission from the particle size distribution measured at the MLO. After a decrease of 10% in the number of particles for different size ranges, the results indicated that fine mode particle change (smaller than diameter size 120 nm) has a small effect on the AOD and transmission measured at the wavelength of 397 - 656 nm. Additionally, the change in the particles having diameters larger than 120 nm change portrayed more AOD and transmission change. Notably, the scale of AOD change estimated from a marine aerosol distribution for particles larger than 120 nm in diameter, as measured at MLO was similar to and larger than the AOD and apparent transmission
change observed by the MOIDS and at MLO.

The DESCAM (DEtailed SCAvenging and Microphysics model) was used to simulate how much aerosols grow by nucleation scavenging, coagulation, diffusion and washout process by cloud processing in terms of time and scale. The ionisation effect due to GCRs change (FD effect) was investigated by the DESCAM simulation. The number size distributions of marine aerosols from Ruprecht Jaenicke 1993 indicated that a 10% decrease in the initial number of aerosols having diameters smaller than 20 nm (FD effect) can cause sufficient enough AOD change measured by wavelengths, 397, 486, 589, and 656 nm after cloud processing, even though the deep convective cloud fraction is small (0.45%). The AOD change after cloud processing was 0.012 (AOD change from the MODIS observation (0.0041±0.0026)). This indicates that cloud processing by deep convective clouds can be a possible factor to explain the observed AOD and cloud optical thickness change due to the change in the number of large aerosols (larger than 120 nm). However, another number size distribution of marine aerosols at MLO did not show sufficient enough AOD change even though the number size of aerosols changes after cloud processing was relatively the same as the result from Ruprecht Jaenicke 1993. Since DESCAM does not include a gas uptake in the simulation, thus we discussed the necessity of this process for future work.