skip to main content

Title: Spatial and temporal variability of snowfall over Greenland from CloudSat observations

Abstract. We use the CloudSat 2006–2016 data record to estimate snowfall over theGreenland Ice Sheet (GrIS). We first evaluate CloudSat snowfall retrievalswith respect to remaining ground-clutter issues. Comparing CloudSatobservations to the GrIS topography (obtained from airborne altimetrymeasurements during IceBridge) we find that at the edges of the GrISspurious high-snowfall retrievals caused by ground clutter occasionallyaffect the operational snowfall product. After correcting for this effect,the height of the lowest valid CloudSat observation is about 1200 mabove the local topography as defined by IceBridge. We then use ground-based millimeter wavelength cloud radar (MMCR) observations obtained from the Integrated Characterization of Energy, Clouds, Atmospheric state, and Precipitation at Summit, Greenland (ICECAPS) experiment to devise a simple,empirical correction to account for precipitation processes occurringbetween the height of the observed CloudSat reflectivities and the snowfallnear the surface. Using the height-corrected, clutter-cleared CloudSatreflectivities we next evaluate various ZS relationships in terms ofsnowfall accumulation at Summit through comparison with weekly stake fieldobservations of snow accumulation available since 2007. Using a set of threeZS relationships that best agree with the observed accumulation at Summit,we then calculate the annual cycle snowfall over the entire GrIS as well asover different drainage areas and compare the derived mean more » values and annualcycles of snowfall to ERA-Interim reanalysis. We find the annual meansnowfall over the GrIS inferred from CloudSat to be 34±7.5 cm yr−1liquid equivalent (where the uncertainty is determined by the range invalues between the three different ZS relationships used). In comparison,the ERA-Interim reanalysis product only yields 30 cm yr−1 liquid equivalentsnowfall, where the majority of the underestimation in the reanalysisappears to occur in the summer months over the higher GrIS and appears to berelated to shallow precipitation events. Comparing all available estimatesof snowfall accumulation at Summit Station, we find the annually averagedliquid equivalent snowfall from the stake field to be between 20 and 24 cm yr−1, depending on the assumed snowpack density and from CloudSat 23±4.5 cm yr−1. The annual cycle at Summit is generally similar betweenall data sources, with the exception of ERA-Interim reanalysis, which showsthe aforementioned underestimation during summer months.

« less
; ; ; ;
Award ID(s):
1801477 1801318
Publication Date:
Journal Name:
Atmospheric Chemistry and Physics
Page Range or eLocation-ID:
8101 to 8121
Sponsoring Org:
National Science Foundation
More Like this
  1. Abstract. Radiation fogs at Summit Station, Greenland (72.58 N,38.48 W; 3210 m a.s.l.), are frequently reported by observers. Thefogs are often accompanied by fogbows, indicating the particles are composedof liquid; and because of the low temperatures at Summit, this liquid issupercooled. Here we analyze the formation of these fogs as well as theirphysical and radiative properties. In situ observations of particle size anddroplet number concentration were made using scattering spectrometers near 2 and 10 m height from 2012 to 2014. These data are complemented bycolocated observations of meteorology, turbulent and radiative fluxes, andremote sensing. We find that liquid fogs occur in all seasons with thehighest frequency in September and a minimum in April. Due to thecharacteristics of the boundary-layer meteorology, the fogs are elevated,forming between 2 and 10 m, and the particles then fall toward the surface.The diameter of mature particles is typically 20–25 µm in summer.Number concentrations are higher at warmer temperatures and, thus, higher insummer compared to winter. The fogs form at temperatures as warm as −5 C, while the coldest form at temperatures approaching −40 C. Facilitated by the elevated condensation, in winter two-thirds offogs occurred within a relatively warm layer above the surface when thenear-surface air was belowmore »−40 C, as cold as −57 C,which is too cold to support liquid water. This implies that fog particlessettling through this layer of cold air freeze in the air column beforecontacting the surface, thereby accumulating at the surface as ice withoutriming. Liquid fogs observed under otherwise clear skies annually imparted1.5 W m−2 of cloud radiative forcing (CRF). While this is a smallcontribution to the surface radiation climatology, individual events areinfluential. The mean CRF during liquid fog events was 26 W m−2, andwas sometimes much higher. An extreme case study was observed toradiatively force 5 C of surface warming during the coldest partof the day, effectively damping the diurnal cycle. At lower elevations ofthe ice sheet where melting is more common, such damping could signal a rolefor fogs in preconditioning the surface for melting later in the day.

    « less
  2. Abstract. Triplet excited states of organic matter are formed when colored organicmatter (i.e., brown carbon) absorbs light. While these “triplets” can beimportant photooxidants in atmospheric drops and particles (e.g., theyrapidly oxidize phenols), very little is known about their reactivity towardmany classes of organic compounds in the atmosphere. Here we measure thebimolecular rate constants of the triplet excited state of benzophenone(3BP), a model species, with 17 water-solubleC3C6 alkenes that have either been found in theatmosphere or are reasonable surrogates for identified species. Measured rateconstants (kALK+3BP) vary by a factor of 30 and are in therange of (0.24–7.5) ×109 M−1 s−1. Biogenic alkenesfound in the atmosphere – e.g., cis-3-hexen-1-ol, cis-3-hexenyl acetate, andmethyl jasmonate – react rapidly, with rate constants above 1×109 M−1 s−1. Rate constants depend on alkene characteristicssuch as the location of the double bond, stereochemistry, and alkylsubstitution on the double bond. There is a reasonable correlation betweenkALK+3BP and the calculated one-electron oxidation potential(OP) of the alkenes (more »class="inline-formula">R2=0.58); in contrast, rate constants are notcorrelated with bond dissociation enthalpies, bond dissociation freeenergies, or computed energy barriers for hydrogen abstraction. Using the OPrelationship, we estimate aqueous rate constants for a number of unsaturatedisoprene and limonene oxidation products with 3BP: values are inthe range of (0.080–1.7) ×109 M−1 s−1, withgenerally faster values for limonene products. Rate constants with lessreactive triplets, which are probably more environmentally relevant, arelikely roughly 25 times slower. Using our predicted rate constants, alongwith values for other reactions from the literature, we conclude thattriplets are probably minor oxidants for isoprene- and limonene-relatedcompounds in cloudy or foggy atmospheres, except in cases in which the tripletsare very reactive.

    « less
  3. Abstract. Secondary organic aerosol derived from isopreneepoxydiols (IEPOX-SOA) is thought to contribute the dominant fraction oftotal isoprene SOA, but the current volatility-based lumped SOAparameterizations are not appropriate to represent the reactive uptake ofIEPOX onto acidified aerosols. A full explicit modeling of this chemistryis however computationally expensive owing to the many species and reactionstracked, which makes it difficult to include it in chemistry–climate modelsfor long-term studies. Here we present three simplified parameterizations(version 1.0) for IEPOX-SOA simulation, based on an approximateanalytical/fitting solution of the IEPOX-SOA yield and formation timescale.The yield and timescale can then be directly calculated using the globalmodel fields of oxidants, NO, aerosol pH and other key properties, and drydeposition rates. The advantage of the proposed parameterizations is thatthey do not require the simulation of the intermediates while retaining thekey physicochemical dependencies. We have implemented the newparameterizations into the GEOS-Chem v11-02-rc chemical transport model,which has two empirical treatments for isoprene SOA (the volatility-basis-set, VBS, approach and a fixed 3 % yield parameterization), and comparedall of them to the case with detailed fully explicit chemistry. The bestparameterization (PAR3) captures the global tropospheric burden of IEPOX-SOAand its spatiotemporal distribution (R2=0.94) vs. thosesimulated by the full chemistry, while being more computationally efficient(∼5 times faster),more »and accurately captures the response tochanges in NOx and SO2 emissions. On the other hand, the constant3 % yield that is now the default in GEOS-Chem deviates strongly (R2=0.66), as does the VBS (R2=0.47, 49 % underestimation), withneither parameterization capturing the response to emission changes. Withthe advent of new mass spectrometry instrumentation, many detailed SOAmechanisms are being developed, which will challenge global and especiallyclimate models with their computational cost. The methods developed in thisstudy can be applied to other SOA pathways, which can allow includingaccurate SOA simulations in climate and global modeling studies in thefuture.

    « less
  4. Abstract. Chemical ionization massspectrometry (CIMS) instruments routinely detect hundreds of oxidized organic compoundsin the atmosphere. A major limitation of these instruments is the uncertaintyin their sensitivity to many of the detected ions. We describe thedevelopment of a new high-resolution time-of-flight chemical ionization massspectrometer that operates in one of two ionization modes: using eitherammonium ion ligand-switching reactions such as for NH4+ CIMS orproton transfer reactions such as for proton-transfer-reaction massspectrometer (PTR-MS). Switching between the modes can be done within 2 min.The NH4+ CIMS mode of the new instrument has sensitivities of upto 67 000 dcps ppbv−1 (duty-cycle-corrected ion counts per second perpart per billion by volume) and detection limits between 1 and 60 pptv at2σ for a 1 s integration time for numerous oxygenated volatileorganic compounds. We present a mass spectrometric voltage scanning procedurebased on collision-induced dissociation that allows us to determine thestability of ammonium-organic ions detected by the NH4+more »xmlns:xlink="" xlink:href="amt-12-1861-2019-ie00005.svg" width="24pt" height="15pt" src="amt-12-1861-2019-ie00005.png"/> CIMS instrument.Using this procedure, we can effectively constrain the sensitivity of theammonia chemical ionization mass spectrometer to a wide range of detectedoxidized volatile organic compounds for which no calibration standards exist.We demonstrate the application of this procedure by quantifying thecomposition of secondary organic aerosols in a series of laboratoryexperiments.

    « less
  5. Abstract. The magnitude and controls of particulate carbon exported from surface watersand its remineralization at depth are poorly constrained. The Carbon FluxExplorer (CFE), a Lagrangian float-deployed imaging sediment trap, has beendesigned to optically measure the hourly variations of particle flux tokilometer depths for months to seasons while relaying data in near-real timeto shore via satellite without attending ships. The main optical proxy forparticle load recorded by the CFE, volume attenuance (VA; units ofmATN cm2), while rigorously defined and highly precise, has not beenrobustly calibrated in terms of particulate organic carbon (POC), nitrogen(PN) and phosphorus (PP). In this study, a novel 3-D-printed particle samplerusing cutting edge additive manufacturing was developed and integrated withthe CFE. Two such modified floats (CFE-Cals) were deployed a total of15 times for 18–24 h periods to gain calibration imagery and samples atdepths near 150 m in four contrasting productivity environments during theJune 2017 California Current Ecosystem Long-Term Ecological Research (LTER)process study. Regression slopes for VA : POC and VA : PN (unitsmATN cm2: mmol; R2, n, p value in parentheses) were1.01×104 (0.86, 12, < 0.001) and 1.01×105(0.86, 15, < 0.001), respectively, and were not sensitive toparticle size classes or the contrasting environments encountered. PP was notwell correlated with VA, reflecting the high labilitymore »of P relative to C andN. The volume attenuance flux (VAF) to POC flux calibration is compared toprevious estimates.

    « less