Search

Cloud microphysics is critical for weather and climate prediction. In this work, we document updates and corrections to the cloud microphysical scheme used in the Community Earth System Model (CESM) and other models. These updates include a new nomenclature for the scheme, now called Parameterization of Unified Microphysics Across Scales (PUMAS), and the ability to run the scheme on graphics processing units (GPUs). The main science changes include refactoring an ice number limiter and associated changes to ice nucleation, adding vapor deposition onto snow, and introducing an implicit numerical treatment for sedimentation. We also detail the improvements in computational performance that can be achieved with GPU acceleration. We then show the impact of these scheme changes on the (a) mean state climate, (b) cloud feedback response to warming, and (c) aerosol forcing. We find that corrections are needed to the immersion freezing parameterization and that ice nucleation has important impacts on climate. We also find that the revised scheme produces less cloud liquid and ice but that this can be adjusted by changing the loss process for cloud liquid (autoconversion). Furthermore, there are few discernible effects of the PUMAS changes on cloud feedbacks but some reductions in the magnitude of aerosol–cloud interactions (ACIs). Small cloud feedback changes appear to be related to the implicit sedimentation scheme, with a number of factors affecting ACIs.

Statistical processing of numerical model output has been a part of both weather forecasting and climate applications for decades. Statistical techniques are used to correct systematic biases in atmospheric model outputs and to represent local effects that are unresolved by the model, referred to as downscaling. Many downscaling techniques have been developed, and it has been difficult to systematically explore the implications of the individual decisions made in the development of downscaling methods. Here we describe a unified framework that enables the user to evaluate multiple decisions made in the methods used to statistically postprocess output from weather and climate models. The Ensemble Generalized Analog Regression Downscaling (En-GARD) method enables the user to select any number of input variables, predictors, mathematical transformations, and combinations for use in parametric or nonparametric downscaling approaches. En-GARD enables explicitly predicting both the probability of event occurrence and the event magnitude. Outputs from En-GARD include errors in model fit, enabling the production of an ensemble of projections through sampling of the probability distributions of each climate variable. We apply En-GARD to regional climate model simulations to evaluate the relative importance of different downscaling method choices on simulations of the current and future climate. We show that choice of predictor variables is the most important decision affecting downscaled future climate outputs, while having little impact on the fidelity of downscaled outcomes for current climate. We also show that weak statistical relationships prevent such approaches from predicting large changes in extreme events on a daily time scale.

The McMurdo Dry Valleys (MDV) are home to a unique microbial ecosystem dependent on the availability of freshwater. It is a polar desert and freshwater originates almost entirely from surface and near-surface melt of cold-based glaciers. Understanding the future evolution of these environments requires the simulation of the full chain of physical processes-from net radiative forcing, surface energy balance, melt, runoff and the transport of meltwater in stream channels from the glaciers to the terminal lakes where the microbial community resides. We present the first application of the WRF-Hydro/Glacier model in the MDV. The model was tested for a 7-month period (1 August 2021 to 28 February 2022) at a point on Commonwealth Glacier and forced by automatic weather station observations. We found it was necessary to limit the percolation of meltwater through ice layers to represent near-surface runoff as observed in the field. We also tuned the parameters controlling the spectral albedo for snow and ice based on observations to model the evolution of broadband albedo over a melt season. With these modifications, we were able to accurately simulate surface and near-surface temperatures, surface height change, broadband albedo and runoff over a melt season. These modifications show that once the model is adapted to this extreme environment, the model is capable of accurately capturing the physical processes governing the meltwater generation of an MDV glacier. This will enable future efforts to model spatially distributed melt and streamflow in the MDV and will allow us to answer questions around the timing of meltwater transport and the present and future hydrological response of melt to atmospheric forcing.

Glacial melt is the primary source of freshwater for the fragile microbial ecosystem in the McMurdo Dry Valleys (MDV) of Antarctica. These glaciers are cold-based, with internal temperatures around -18°C, however, air temperatures hover around 0°C for several weeks in the summer and föhn wind events can rapidly raise ice surface temperatures to the melting point. Thus, episodical glacial melt is sensitive to small changes in the climate. The aim of this research is to adapt a detailed snowpack model embedded in a distributed hydrological model to simulate the surface energy balance and run-off of a glacier in the MDV. To do this, the snowpack model in the WRF-Hydro-Crocus modelling scheme, which has been used for avalanche forecasting and temperate glaciers, is adapted to the MDV. Several modifications are made to model calculations and parameters to allow the model to successfully simulate surface energy balance and runoff in this environment. For example, the parameters for the Crocus albedo scheme are adjusted to obtain band profiles for snow, firn and ice that replicate observed albedo and remain internally consistent between surface types. The modelling system is then validated against data from an automatic weather station, eddy covariance measurements and stream discharge. It is shown to be suitable for future efforts to model the full hydrological cycle of glacial meltwater in this region.

Occurrence frequency and dynamical conditions of ice supersaturation (ISS, where relative humidity with respect to ice (RHi) greater than 100%) are examined in the upper troposphere around convective activity. Comparisons are conducted between in situ airborne observations and the Weather Research and Forecasting model simulations using four double-moment microphysical schemes at temperatures less than or or equal to -40degdegC. All four schemes capture both clear-sky and in-cloud ISS conditions. However, the clear-sky (in-cloud) ISS conditions are completely (significantly) limited to the RHi thresholds of the Cooper parameterization. In all of the simulations, ISS occurrence frequencies are higher by approximately 3-4 orders of magnitude at higher updraft speeds (greater than 1 m s(exp -1) than those at the lower updraft speeds when ice water content (IWC) greater than 0.01 gm(exp -3), while observations show smaller differences up to approximately 1-2 orders of magnitude. The simulated ISS also occurs less frequently at weaker updrafts and downdrafts than observed. These results indicate that the simulations have a greater dependence on stronger updrafts to maintain/generate ISS at higher IWC. At lower IWC (less than or equal or 0.01 gm(exp -3), simulations unexpectedly show lower ISS frequencies at stronger updrafts. Overall, the Thompson aerosol-aware scheme has the closest magnitudes and frequencies of ISS greater than 20% to the observations, and the modified Morrison has the closest correlations between ISS frequencies and vertical velocity at higher IWC and number density. The Cooper parameterization often generates excessive ice crystals and therefore suppresses the frequency and magnitude of ISS, indicating that it should be initiated at higher ISS (e.g.,lees than or equal to 25%).

This paper describes a new approach for representing ice microphysics in climate models. In contrast with most previous schemes, this approach does not include separate categories for cloud and precipitating ice and instead uses a single two-moment category to represent all solid hydrometeors. Thus, there is no need for an ice “autoconversion” size threshold parameter, which has a critical impact on simulated climate in the Community Atmosphere Model (CAM5) yet is poorly constrained by theory or observations. Further, in the new treatment, all ice microphysical processes and parameters, including ice effective radius and mean fall speed, are formulated self-consistently and flexibly based on empirical ice particle mass–size and projected area–size relationships. This means that the scheme can represent the physical coupling between bulk particle density, mean fall speed, and effective radius, which is not possible in current schemes. Two different methods for specifying these relationships based on observations are proposed. The new scheme is tested in global simulations using CAM5. Differences in simulations using the two methods for specifying the mass– and projected area–size relationships, particularly the cloud radiative forcing, are attributable mainly to the effects on mean ice particle fall speed, impacting sedimentation and ice water path. With some tuning of parameters involved in calculating homogeneous freezing it produces a similar climate compared to the simulations using the original CAM5 microphysics. Thus, it can produce a comparable climate while improving the physical basis and self-consistency of ice particle properties and parameters.

Orographic precipitation and snowpack provide a vital water resource for the western U.S., while convective precipitation accounts for a significant part of annual precipitation in the eastern U.S. As a result, water managers are keenly interested in their fate under climate change. However, previous studies of water cycle changes in the U.S. have been conducted with climate models of relatively coarse resolution, leading to potential misrepresentation of key physical processes. This paper presents results from a high-resolution climate change simulation that permits convection and resolves mesoscale orography at 4-km grid spacing over much of North America using the Weather Research and Forecasting (WRF) model. Two 13-year simulations were performed, consisting of a retrospective simulation (October 2000–September 2013) with initial and boundary conditions from ERA-interim and a future climate sensitivity simulation with modified reanalysis-derived initial and boundary conditions through adding the CMIP5 ensemble-mean high-end emission scenario climate change. The retrospective simulation is evaluated by validating against Snowpack Telemetry (SNOTEL) and an ensemble of gridded observational datasets. It shows overall good performance capturing the annual/seasonal/sub-seasonal precipitation and surface temperature climatology except for a summer dry and warm bias in the central U.S. In particular, the WRF seasonal precipitation agrees with SNOTEL observations within a few percent over the mountain ranges, providing confidence in the model’s estimation of western U.S. seasonal snowfall and snowpack. The future climate simulation forced with warmer and moister perturbed boundary conditions enhances annual and winter-spring-fall seasonal precipitation over most of the contiguous United States (CONUS), but suppresses summertime precipitation in the central U.S. The WRF-downscaled climate change simulations provide a high-resolution dataset (i.e., High-Resolution CONUS downscaling, HRCONUS) to the community for studying one possible scenario of regional climate changes and impacts.

WRF-Solar is a specific configuration and augmentation of the Weather Research and Forecasting (WRF) Model designed for solar energy applications. Recent upgrades to the WRF Model contribute to making the model appropriate for solar power forecasting and comprise 1) developments to diagnose internally relevant atmospheric parameters required by the solar industry, 2) improved representation of aerosol–radiation feedback, 3) incorporation of cloud–aerosol interactions, and 4) improved cloud–radiation feedback. The WRF-Solar developments are presented together with a comprehensive characterization of the model performance for forecasting during clear skies. Performance is evaluated with numerical experiments using a range of different external and internal treatment of the atmospheric aerosols, including both a model-derived climatology of aerosol optical depth and temporally evolving aerosol optical properties from reanalysis products. The necessity of incorporating the influence of atmospheric aerosols to obtain accurate estimations of the surface shortwave irradiance components in clear-sky conditions is evident. Improvements of up to 58%, 76%, and 83% are found in global horizontal irradiance, direct normal irradiance, and diffuse irradiance, respectively, compared to a standard version of the WRF Model. Results demonstrate that the WRF-Solar model is an improved numerical tool for research and applications in support of solar energy.

We use the WRF system to study the impacts of biomass burning smoke from Central America on several tornado outbreaks occurring in the US during spring. The model is configured with an aerosol-aware microphysics parameterization capable of resolving aerosol-cloud-radiation interactions in a cost-efficient way for numerical weather prediction (NWP) applications. Primary aerosol emissions are included and smoke emissions are constrained using an inverse modeling technique and satellite-based AOD observations. Simulations turning on and off fire emissions reveal smoke presence in all tornado outbreaks being studied and show an increase in aerosol number concentrations due to smoke. However, the likelihood of occurrence and intensification of tornadoes is higher due to smoke only in cases where cloud droplet number concentration in low level clouds increases considerably in a way that modifies the environmental conditions where the tornadoes are formed (shallower cloud bases and higher low-level wind shear). Smoke absorption and vertical extent also play a role, with smoke absorption at cloud-level tending to burn-off clouds and smoke absorption above clouds resulting in an increased capping inversion. Comparing these and WRF-Chem simulations configured with a more complex representation of aerosol size and composition and different optical properties, microphysics and activation schemes, we find similarities in terms of the simulated aerosol optical depths and aerosol impacts on near-storm environments. This provides reliability on the aerosol-aware microphysics scheme as a less computationally expensive alternative to WRF-Chem for its use in applications such as NWP and cloud-resolving simulations.

We present an aerosol-cloud-precipitation modeling study of convective clouds using the Weather Research and Forecasting model fully coupled with Chemistry (WRF-Chem) version 3.1.1. Comparison of the model output with measurements from a research site in the Rocky Mountains in Colorado revealed that the fraction of organics in the model is underpredicted. This is most likely due to missing processes in the aerosol module in the model version used, such as new particle formation and growth of secondary organic aerosols. When boundary conditions and domain-wide initial conditions of aerosol loading are changed in the model (factors of 0.1, 0.2, and 10 of initial aerosol mass of SO4−2, NH4+, and NO3−), the domain-wide precipitation changes by about 5%. Analysis of the model results reveals that the Rocky Mountain region and Front Range environment is not conducive for convective invigoration to play a major role, in increasing precipitation, as seen in some other studies. When localized organic aerosol emission are increased to mimic new particle formation, the resulting increased aerosol loading leads to increases in domain-wide precipitation, opposite to what is seen in the model simulations with changed boundary and initial conditions.

Aerosols influence cloud and precipitation development in complex ways due to myriad feedbacks at a variety of scales from individual clouds through entire storm systems. This paper describes the implementation, testing, and results of a newly modified bulk microphysical parameterization with explicit cloud droplet nucleation and ice activation by aerosols. Idealized tests and a high-resolution, convection-permitting, continental-scale, 72-h simulation with five sensitivity experiments showed that increased aerosol number concentration results in more numerous cloud droplets of overall smaller size and delays precipitation development. Furthermore, the smaller droplet sizes cause the expected increased cloud albedo effect and more subtle longwave radiation effects. Although increased aerosols generally hindered the warm-rain processes, regions of mixed-phase clouds were impacted in slightly unexpected ways with more precipitation falling north of a synoptic-scale warm front. Aerosol impacts to regions of light precipitation, less than approximately 2.5 mm h−1, were far greater than impacts to regions with higher precipitation rates. Comparisons of model forecasts with five different aerosol states versus surface precipitation measurements revealed that even a large-scale storm system with nearly a thousand observing locations did not indicate which experiment produced a more correct final forecast, indicating a need for far longer-duration simulations due to the magnitude of both model forecast error and observational uncertainty. Last, since aerosols affect cloud and precipitation phase and amount, there are resulting implications to a variety of end-user applications such as surface sensible weather and aircraft icing.

Detailed measurements of ice crystals in cirrus clouds were used to compare with results from the Community Atmospheric Model Version 5 (CAM5) global climate model. The observations are from two different field campaigns with contrasting conditions: Atmospheric Radiation Measurements Spring Cloud Intensive Operational Period in 2000 (ARM-IOP), which was characterized primarily by midlatitude frontal clouds and cirrus, and Tropical Composition, Cloud and Climate Coupling (TC4), which was dominated by anvil cirrus. Results show that the model typically overestimates the slope parameter of the exponential size distributions of cloud ice and snow, while the variation with temperature (height) is comparable. The model also overestimates the ice/snow number concentration (0th moment of the size distribution) and underestimates higher moments (2nd through 5th), but compares well with observations for the 1st moment. Overall the model shows better agreement with observations for TC4 than for ARM-IOP in regards to the moments. The mass-weighted terminal fall speed is lower in the model compared to observations for both ARM-IOP and TC4, which is partly due to the overestimation of the size distribution slope parameter. Sensitivity tests with modification of the threshold size for cloud ice to snow autoconversion (Dcs) do not show noticeable improvement in modeled moments, slope parameter and mass weighed fall speed compared to observations. Further, there is considerable sensitivity of the cloud radiative forcing to Dcs, consistent with previous studies, but no value of Dcs improves modeled cloud radiative forcing compared to measurements. Since the autoconversion of cloud ice to snow using the threshold size Dcs has little physical basis, future improvement to combine cloud ice and snow into a single category, eliminating the need for autoconversion, is suggested.

Meltwater runoff in the catchment area containing Chhota Shigri glacier (Western Himalaya) is simulated for the period 1951–2099. The applied mass-balance model is forced by downscaled products from four regional climate models with different horizontal resolution. For the future climate scenarios we use high resolution time series of 5 km grid spacing, generated using the newly developed Intermediate Complexity Atmospheric Research Model. The meteorological input is downscaled to 300 m horizontal resolution. The use of an ice flow model provides annually updated glacier area for the mass-balance calculations. The mass-balance model calculates daily snow accumulation, melt, runoff, as well as the individual runoff components (glacial melt, snowmelt and rain). The resulting glacier area decreases by 35% (representative concentration pathway (RCP) 4.5 scenario) to 70% (RCP 8.5 scenario) by 2099 relative to 2000. The average annual mass balance over the whole model period (1951–2099) was –0.4 (±0.3) m w.e. a–1. Average annual runoff does not differ substantially between the two climate scenarios. However, for the years after 2040 our results show a shift towards earlier snowmelt onset that increases runoff in May and June, and reduced glacier melt that decreases runoff in August and September. This shift is much stronger pronounced in the RCP 8.5 scenario.

Glacier mass balance and runoff are simulated from 1955 to 2014 for the catchment (46% glacier cover) containing Chhota Shigri Glacier (Western Himalaya) using gridded data from three regional climate models: (1) the Rossby Centre regional atmospheric climate model v.4 (RCA4); (2) the REgional atmosphere MOdel (REMO); and (3) the Weather Research and Forecasting Model (WRF). The input data are downscaled to the simulation grid (300 m) and calibrated with point measurements of temperature and precipitation. Additional input is daily potential global radiation calculated using a DEM at a resolution of 30 m. The mass-balance model calculates daily snow accumulation, melt and runoff. The model parameters are calibrated with available mass-balance measurements and results are validated with geodetic measurements, other mass-balance model results and run-off measurements. Simulated annual mass balances slightly decreased from −0.3 m w.e. a−1 (1955–99) to −0.6 m w.e. a−1 for 2000–14. For the same periods, mean runoff increased from 2.0 m3 s−1 (1955–99) to 2.4 m3 s−1 (2000–14) with glacier melt contributing about one-third to the runoff. Monthly runoff increases are greatest in July, due to both increased snow and glacier melt, whereas slightly decreased snowmelt in August and September was more than compensated by increased glacier melt.

Knowledge of cloud and precipitation formation processes remains incomplete, yet global precipitation is predominantly produced by clouds containing the ice phase. Ice first forms in clouds warmer than -36 °C on particles termed ice nuclei. We combine observations from field studies over a 14-year period, from a variety of locations around the globe, to show that the concentrations of ice nuclei active in mixed-phase cloud conditions can be related to temperature and the number concentrations of particles larger than 0.5 μm in diameter. This new relationship reduces unexplained variability in ice nuclei concentrations at a given temperature from ∼10 3 to less than a factor of 10, with the remaining variability apparently due to variations in aerosol chemical composition or other factors. When implemented in a global climate model, the new parameterization strongly alters cloud liquid and ice water distributions compared to the simple, temperature-only parameterizations currently widely used. The revised treatment indicates a global net cloud radiative forcing increase of ∼1 W m -2 for each order of magnitude increase in ice nuclei concentrations, demonstrating the strong sensitivity of climate simulations to assumptions regarding the initiation of cloud glaciation.

Detailed measurements of ice crystals in cirrus clouds were used to compare with results from the Community Atmospheric Model Version 5 (CAM5) global climate model. The observations are from two different field campaigns with contrasting conditions: Atmospheric Radiation Measurements Spring Cloud Intensive Operational Period in 2000 (ARM-IOP), which was characterized primarily by midlatitude frontal clouds and cirrus, and Tropical Composition, Cloud and Climate Coupling (TC4), which was dominated by anvil cirrus. Results show that the model typically overestimates the slope parameter of the exponential size distributions of cloud ice and snow, while the variation with temperature (height) is comparable. The model also overestimates the ice/snow number concentration (0th moment of the size distribution) and underestimates higher moments (2nd through 5th), but compares well with observations for the 1st moment. Overall the model shows better agreement with observations for TC4 than for ARM-IOP in regards to the moments. The mass-weighted terminal fallspeed is lower in the model compared to observations for both ARM-IOP and TC4, which is partly due to the overestimation of the size distribution slope parameter. Sensitivity tests with modification of the threshold size for cloud ice to snow autoconversion (Dcs) do not show noticeable improvement in modeled moments, slope parameter and mass weighed fallspeed compared to observations. Further, there is considerable sensitivity of the cloud radiative forcing to Dcs, consistent with previous studies, but no value of Dcs improves modeled cloud radiative forcing compared to measurements. Since the autoconversion of cloud ice to snow using the threshold size Dcs has little physical basis, future improvement to combine cloud ice and snow into a single category, eliminating the need for autoconversion, is suggested.

The temperature of dust in the Earth's mesosphere, which is determined by collisions with the neutral gas, absorption of radiation from the Sun, the Earth and its own thermal radiation, in general increases with dust size [Grams and Fiocco, 1977]. This may lead to situations where there is a maximum size of the dust rC above which the dust temperature is too high for condensation of water vapor to occur. Dust of this size can still accrete other elements. If they accrete a sufficient amount of elements, with low ionization potential as a surface “contamination,” their photoelectric properties may change. We suggest that this can lead to their charges changing from the “normal” low negative value of pure ice particles to the positive charges which rocket in situ observations have shown to exist. In such cases we will have a situation with small, newly created and negatively charged dust particles coexisting with larger and positively charged dust charges. Agglomeration of dust particles will then be an effective process. We also demonstrate that the changes in the polar surface temperature of the Earth during the polar mesosphere summer echo (PMSE) season can influence the mesospheric dust temperature and by this have a potential effect on the shape of the occurrence rate curve of the PMSE.

The initiation of ice in an isolated orographic wave cloud was compared with expectations based on ice nucleating aerosol concentrations and with predictions from new ice nucleation parameterizations applied in a cloud parcel model. Measurements of ice crystal number concentrations were found to be in good agreement both with measured number concentrations of ice nuclei feeding the clouds and with ice nuclei number concentrations determined from the residual nuclei of cloud particles collected by a counterflow virtual impactor. Using lognormal distributions fitted to measured aerosol size distributions and measured aerosol chemical compositions, ice nuclei and ice crystal concentrations in the wave cloud were reasonably well predicted in a 1D parcel model framework. Two different empirical parameterizations were used in the parcel model: a parameterization based on aerosol chemical type and surface area and a parameterization that links ice nuclei number concentrations to the number concentrations of particles with diameters larger than 0.5 μm. This study shows that aerosol size distribution and composition measurements can be used to constrain ice initiation by primary nucleation in models. The data and model results also suggest the likelihood that the dust particle mode of the aerosol size distribution controls the number concentrations of the heterogeneous ice nuclei, at least for the lower temperatures examined in this case.

Variations in the chemical composition of atmospheric aerosols alter their hygroscopicity and can lead to changes in the cloud-active fraction of the aerosols, or cloud condensation nuclei (CCN) number concentration. To investigate the importance of this effect under different atmospheric conditions, cloud droplet formation was simulated with a Lagrangian parcel model. Initial values of updraft speed and temperature were systematically varied along with aerosol number concentration, size and hygroscopicity (represented by the hygroscopicity parameter, κ). A previous study classifies the sensitivity of CCN activity to compositional changes based on the supersaturation reached in the parcel model. We found that these classifications could not be generalized to a range of aerosol size distribution median radii. Instead, variations in sensitivity with size depend on the location of the dry critical radius for droplet activation relative to the size distribution median radius. The parcel model output was used to construct droplet activation lookup tables based on κ that were implemented in the Regional Atmospheric Modeling System (RAMS) microphysical scheme. As a first application of this system, aerosol hygroscopicity and size were varied in a series of RAMS mesoscale simulations designed to investigate the sensitivity of a mixed-phase orographic cloud case to the parameter variations. Observations from a recent field campaign in northwestern Colorado provided the basis for the aerosol field initializations. Model results show moderate sensitivity in the distribution of total case precipitation to extreme changes in κ, and minimal sensitivity to observed changes in estimated κ. The impact of varying aerosol hygroscopicity diminished with increasing median radius, as expected from the parcel model results. The conclusions drawn from these simulations could simplify similar research in other cloud regimes by defining the need, or lack of need, for detailed knowledge of aerosol composition.

[1] A liquid-phase Lagrangian parcel model was expanded to include nucleation and growth of ice crystals. Intercomparisons between three heterogeneous ice nucleation parameterizations that link aerosol type and number to ice crystal concentration were conducted. Results indicate large differences in the prediction of ice formation in modestly supercooled clouds and in the susceptibility of cirrus to heterogeneous ice nucleation for the same assumed aerosol distribution. Only one parameterization has an observational constraint that limits the maximum ice crystal number concentrations to be a fraction of the total number concentrations of potential ice-nucleating particles (typically, all insoluble particles larger than about 0.1 μm). The constrained parameterization compares well with most ice nucleation measurements. The nonconstrained parameterizations are capable of predicting several orders of magnitude higher ice crystal concentrations than the constrained parameterization for the same parcel-forcing conditions. Ice crystal concentrations in the unconstrained parameterizations are controlled by the total number concentration of potential ice-nucleating particles and, importantly, negative feedback on ice supersaturation. This feedback control often masks the large discrepancy that exists between predicted ice crystal number concentrations and the maximum number concentrations that can be attributed to atmospheric ice nuclei based on our current understanding of the latter. It also permits unrealistic conclusions regarding the role of certain aerosols as ice nuclei. It is recommended that a constraint on ice crystal number concentrations, related to number concentrations of relevant aerosol particles, should be included in ice nucleation parameterizations used in cloud to global-scale models.

[1] Many soil-derived particles dominated by insoluble material, including Saharan dusts, are known to act as ice nuclei. If, however, dust particles can compete with other atmospheric particle types to form liquid cloud droplets, they have a greater potential to change climate through indirect effects on cloud radiative properties and to affect the hydrological cycle through precipitation changes. By directly collecting and analyzing the residual nuclei of small cloud droplets, we demonstrate that Saharan dust particles do commonly act as cloud condensation nuclei (CCN) in the eastern North Atlantic. Droplet activation calculations support the measurements by showing that due to its slightly hygroscopic nature, even submicron dust can be important as CCN. Given the dual nature of Saharan dust particles as CCN and ice nuclei, this infusion of dust is expected to impact not only droplet size and albedo in small clouds, but ice formation in deep convective clouds.

[1] A twin angle optical particle counter (TAOPC) that measures forward scattering at 40° and 74° was developed to determine the index of refraction of atmospheric particles. An error analysis indicates that measurement uncertainties for size at 40° are between 4 and 10% for particles >1.5 μm and between 5 and 15% at 74°. For the index of refraction determination, the measurement uncertainties lead to index of refraction uncertainties between 1 and 2.5%. The instrument was tested on nonabsorbing spherical particles of known composition and size in the laboratory. The majority of the estimated indices of refraction were within ±1% of the expected indices, and size determination was within acceptable error. The instrument was also tested on non-spherical absorbing particles to determine the complex index of refraction of ambient mineral dust particles collected in Laramie, Wyoming, in February 2006. The index of refraction was determined with the particle number ratio approach and was estimated to be in the range 1.60–1.67 for the real part and 0.009–0.0104 for the imaginary part. Simultaneously with the TAOPC measurements, particles were collected on polycarbonate filters for computer-controlled scanning electron microscopy (CCSEM) analysis. Index of refraction calculated from this analysis was in the range 1.61–1.66 for the real part and 0.008–0.012 for the imaginary part. Particles were also collected for longer periods on two different filter pack systems in February 2006. Estimates of index of refraction from these measurements compared well with the CCSEM analysis.