Boundary layer mixing heights and uncalibrated 2 m

Boundary layer mixing heights and uncalibrated 2 µm backscatter distributions measured during Tex. AQS 2006 Sara C. Tucker*, Alan Brewer, Christoph Senff*, Robert Banta, Mike Hardesty, Scott Sandberg, Ann Weickmann*, Dan Law Optical Remote Sensing Group, Chemical Sciences Division (CSD) Earth System Research Laboratory, NOAA *Also with: Cooperative Institute for Research in Environmental Science University of Colorado, Boulder, CO, NOAA/ESRL/CSD Lidar Working Group: February 5 -8, 2008, Monterey, CA Working Group on Space-Based Lidar Winds, Monterrey, CA February 4 -7, 2008

HIGH RESOLUTION DOPPLER LIDAR (HRDL) 0 4 min conical scans (1, 7 and 45 degree elevations) 6 min Zenith Stare 5 min Elevation Scans 15 45 HRDL scan sequence 30 RV Brown HRDL Scan sequence repeats every 15 minutes HRDL Parameters Wavelength 2. 02 µm Pulse Energy 2 -3 m. J Pulse Repetition Frequency 200 Hz Maximum Range 4 -8 km Range Resolution 30 m Beam rate 2 Hz Scanning Full Hemispheric Precision 15 -30 cm/s

HRDL analysis – what’s new? p Mixing Height estimation is finished and results now being used for: n Mixing Ratios (e. g. for secondary formaldehyde production) n Small scale WRF modeling validation n Atmospheric scientist “buy-in” for Doppler lidar products p Analysis of the effects of the relationship of the nocturnal LLJ to Houston ozone levels. p Statistics on relative aerosol backscatter profiles (Size distribution data not yet available to generate actual 2 micron backscatter).

Lidar profiles and methods for 15 minute MIXING HEIGHT ESTIMATION p hσ from composite velocity variance σV 2 profiles (horizontal σH 2 and vertical σw 2 ) p h. A from profiles of relative aerosol backscatter SRAB p hspd, hdir, and hvec from horizontal mean wind profiles

MIXING HEIGHT PRODUCT p Boundary layer: That layer of the atmosphere in turbulent connection with the surface of the earth, on time scales of about an hour (Stull, 1988) p hσ, hspd, hdir, hvec, and h. A are all candidates for MH – but hσ is the most appropriate since it directly represents turbulence. p The others are related to turbulence as the cause (shear) or effect (aerosol gradient) and so may be used when hσ is not available.

Vertical velocity variance σw 2 profiles are calculated using vertical velocity data from each vertical stare period. These can then be displayed using color for amplitude.

Mixing or Mixed? The top of the aerosol layer does not always correspond to the mixing height.

Data from HRDL elevation scans are binned and used to calculate estimates of horizontal variance σH 2 that comprise the lowest 300 meters of the σV 2 profile Sonde profiles indicate a 200 m mixed layer Horizontal velocity variance drops below threshold at the same altitude as the wind speed shear profile.

Full Mixing Profiles

Product used for Mixing Height

Mixing Heights by Hour/Day

HRDL vs. Sonde MH

Mixing Height Averages

WRF model validation Gulf of Mexico and Galveston Bay

MIXING HEIGHT ESTIMATION: Summary (For Tex. AQS 2006 data) p High variability over land – doesn’t match aerosol profiles, p Less variability (in mixing or aerosol layers) over Gulf p More clouds over Gulf p Velocity variance σV 2 profiles used here for MH later for TKE to feed into models

Sunset Sunrise Log scale intensity Turbulence follows a diurnal cycle due to land influence Turbulence is fairly constant, until late afternoon at which point the ship is near the shipping lanes outside of Galveston. Nocturnal Low level jet Speed shear Northerly and Northwesterly flow aloft South-Southwesterly flow South/Southeasterly flow Sea breeze Southeasterly

HRDL-Tex. AQS 2006: Zenith Return Signal Strength

Distributions in relative aerosol backscatter (above threshold) Inland

Distributions in relative aerosol backscatter (above threshold) Gulf of Mexico

Backscatter Inversion p p Can’t pull signal out from below noise floor. CNR mismatch Refractive turbulence effects. Changes in noise floor: clouds in noisebins, RF noise, etc. 5000 4000 3000 2000 1000 0 -20 -15 -10 -5 0 5 10 15 20

Next… p p p Implement more thorough CNR filtering and fitting algorithms for backscatter inversion Integrate size distribution data into Mie backscatter model. Compare HRDL data with in-situ-based calculations of backscatter Variance profiles TKE profiles: integrate into CNR models Quantify the effects of these conditions on space-based lidar CNR & wind estimates.

Particle size distributions: concentrations and backscatter Likely hard target returns

$2µm Backscatter: Caveats p CNR fit depends on: n n p p p Refractive$

2µm Backscatter: Caveats p CNR fit depends on: n n p p p Refractive turbulence Transmission/extinction (estimated in Mie models) Ship plume – strong refractive turbulence Possible long-term system changes due to high vibration Mie scattering models still “young” n Particle refractive index is highly composition dependent): Incorporate variable mass fractions of Ammonium Sulfate, Sea Salt, and Dust

Relative aerosol backscatter profiles

Distributions in relative aerosol backscatter (above threshold) Inland

Distributions in relative aerosol backscatter (above threshold) Gulf of Mexico

Lidar profiles and methods for 15 minute MIXING HEIGHT ESTIMATION n hσ - Find the height at which surface-connected velocity variance falls below 0. 04 m 2 s-2. n h. RAB - Use the Haar wavelet method to detect gradient at top of surface-based aerosol layer. n hspd - If speed shear greater than 0. 05 s-1 is present, find the altitude at which the shear falls below 0. 01 s-1. n hdir - If directional shear greater than 0. 05 °m-1 is present, find the altitude at which the shear falls below 0. 01 °m-1. n hvec - If vector shear greater than 2 s-1 is present, find the altitude at which the shear falls below 0. 01 s-1.

Estimating Velocity Variance

$Mass fractions of Ammonium Sulfate, Sea Salt, and Dust$

Mass fractions of Ammonium Sulfate, Sea Salt, and Dust

Accuracy and Precision Low accuracy, high precision High accuracy, low precision p p p Accuracy – how far off is the mean? Bias. Precision – what is the standard deviation of the measurements? Averaging more ACFs or Spectra typically means better precision – but may not mean the results are accurate.