GOESR3 Periodic Reporting
Project Team: Xiquan Dong, Zhanqing Li, Bob Kuligowski
Reporting Period: February 2013-June 2014
Team Lead: Xiquan Dong
Team Members: Ronald Stenz, Jingyu Wang, and Ning Zhou (Dong’s graduate students), Zhanqing Li,
Bob Kuligowski
Project Title: Improving GOES-R Cloud and Precipitation Products Associated with Deep Convective
Systems by using NEXRAD Radar Network over the Continental U.S.
Project Number:
86
Executive Summary
One of the GOES-R goals is to improve operational satellite-based cloud and precipitation products
to enhance short-term heavy rainfall and flood forecast, as well as long-term assessments concerning
agriculture and water resources management. Most heavy precipitation are associated with deep
convective systems (DCSs), whose large-scale morphologic feature of a cold cloud shield at the
tropopause-level and cloud microphysical properties (phase, size, LWP, etc.) near cloud top can be
monitored by the GOES-R. It is difficult, however, to separate precipitating portions of a DCS from
non-precipitating anvils from GOES-R observations, which leads to large uncertainties in satellite
IR-based precipitation retrievals. This key limitation can be improved by using a newly developed
automatic 3-D radar (NEXRAD) classification technique to identify the convective and stratiform
rain regions (precipitation) and cirrus anvil regions (non-precipitating) from midlatitude DCSs [Feng
et al. 2011]. By integrating the radar classification technique with satellite cloud property retrievals,
we are developing a satellite-microphysics-based cloud classification to improve the current IRbased precipitation retrieval algorithm, and then evaluating its improvement against ground-based
radar observations and aircraft in situ data.
Four proposed research objectives:
(1) to use the newly developed classification technique to identify DCSs and classify them into
convective core, stratiform rain region, mixed and ice anvils, and then to obtain GOES-R
cloud and precipitation product statistics for each classified region over the continental U.S.,
(2) to use the multi-scale, multi-sensor ground-based radars and other observations, and
aircraft in situ measurements to validate satellite retrieved cloud and precipitation properties
from the Mid-latitude Convective Clouds Experiment (MC3E) during the late spring/early
summer of 2011 within the ARM SGP site,
(3) to investigate the impact of aerosols on DCSs. Aerosols may affect the development of a
DCS by delaying precipitation, enhancing latent heat release and invigorating clouds to
become severe storms. It is unclear to what extent and under what conditions such effects
occur. By using rich information from multiple satellite datasets, we are investigating these
effects as a pathway for the exploitation of future GOES-R products.
(4) to take advantage of new satellite and ground-based remote sensing cloud products to test,
improve, and develop cloud parameterization schemes used in the NCEP/GFS model. This
may lay a foundation for future use of GOES-R products in operational weather applications.
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R
Milestones
1) Have collected all NEXRAD radar and GOES data over the continental U.S. to generate the
hybrid classification product and to identify the raincore and non-raincore regions of DCS. We
also have collected NEXRAD Q2/Q3 precipitation products.
2) Have quantitatively evaluated SCaMPR precipitation retrievals using Oklahoma MESONET and
NEXRAD Q2 precipitation. The paper has been accepted by J of Hydrometeor (Mr. Stenz and
Dr. Dong at UND).
3) Have investigated the DCS microphysical properties and precipitation characteristics using the
UND citation research aircraft and DOE ARM and NOAA radars observations. Two papers have
been submitted to JGR from Dr. Dong’s group (Tian et al. 2014, and Wang et al. 2014).
4) Will develop a new method to improve the SCaMPR precipitation rate estimates during the
period 2014-2015. (Ron Stenz and Dong at UND).
5) Have developed an automatic algorithm to identify DCSs using A-train satellite data to obtain
basic statistics of DCSs around the globe, which can be used as a benchmark for testing similar
algorithms for GOES-R (Dr. Li at UMD)
6) Have investigated the impact of aerosols on DCSs by analyzing multiple years of matched data
between CloudSat, CALIPSO, MODIS. A past study using one year of data showed that aerosols
invigorate DCSs by increasing cloud top height and cloud thickness, (increase/decrease ?)
rainfall frequency substantially, but have little impact on low-level liquid clouds. The study will
be enriched by evaluating many more data samples (Dr. Li at UMD)
7) Use DCSs identified by satellite-based algorithms to assess the performance of NCEP/GFS
model in treating DCSs. Our earlier study (Yoo et al. 2012) demonstrated that these products are
highly valuable in helping diagnosis model deficiencies. In 2013, we will consider different
types of clouds, especially low stratus clouds and DCSs (Dr. Li at UMD)
Accomplishments & Plans
Section A: Dr. Xiquan Dong’s group at University of North Dakota
2.1. Dr. Dong’s student, Mr. Ronald Stenz, submitted (accepted) a paper to the Journal of
Hydrometeorology. They assessed the performance of NEXRAD Q2 precipitation and SCaMPR
precipitation retrievals over the state of Oklahoma (OK) using OK MESONET observations as ground
truth. While the average annual Q2 precipitation estimates were about 35% higher than MESONET
observations (~690 mm), there were very strong correlations between these two data sets for multiple
temporal and spatial scales. SCaMPR retrievals were typically three to four times higher than the
collocated MESONET observations, with relatively weak correlations to OK MESONET observations
during 2012. The severe overestimations from SCaMPR retrievals were primarily attributed to false
alarm retrieval of heavy precipitation in anvil regions during DCS events.
Stenz, R., X. Dong, B. Xi, and B. Kuligowski (2014): Assessment of SCaMPR and NEXRAD Q2
precipitation estimates using Oklahoma Mesonet observations. Accepted by J. Hydrometeor.
2.2. Dr. Dong’s MS student, Mr. Ronald Stenz has developed a satellite-microphysics-based cloud
and precipitation classification algorithm using GOES cloud optical depth and IR temperature
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R
As mentioned above, the SCaMPR retrieval method is based on satellite IR temperature, which makes it
difficult to separate precipitating portions of DCSs from non-precipitating anvils using GOES-R
observations, leading to large uncertainties in satellite IR-based precipitation retrievals.
A satellite-microphysics-based cloud and precipitation classification algorithm has been developed over
the SGP region [Stenz et al. 2014] to qualitatively improve the SCaMPR precipitation estimates by
utilizing a strong optical depth gradient that was found between the precipitating and non-precipitating
(anvil) regions of DCSs. This newly developed precipitation mask can identify the areal precipitation
coverage using GOES cloud optical depth and IR BT, instead of IR BT only as in the SCaMPR method.
This improved precipitation product would be a valuable asset for NWS applications.
A modified SCaMPR retrieval algorithm, employing both cloud optical depth and IR temperature, has
made significant improvements to reduce the SCaMPR false alarm rate of retrieved precipitation
especially over non-precipitating (anvil) regions of a DCS. As illustrated in Fig. 1, using the newly
developed satellite-microphysics-based cloud classification, the spatial extent of the SCaMPR estimated
precipitation was reduced to 31% in the modified version from 48% in its original algorithm (IR
temperature only). The new coverage is very close to the Q2 estimated precipitation coverage (33%,
Fig. 1d). A more robust comparison covering 14 convective events during the Midlatitude Continental
Convective Clouds Experiment (MC3E) campaign at the ARM SGP site has also shown the
precipitation area estimated from the modified algorithm (9.64%) is closer to the Q2 estimation
(12.06%) than that (19.11%) from the SCaMPR original algorithm.
Additional analysis of the new algorithm for 16 days with widespread convective activity during
2012 shows a significant improvement in the distribution of the amount of precipitation among the DCS
components. After incorporating RH and applying the new rain mask, over 70% of estimated
precipitation from SCaMPR retrievals occurred in raincore regions, while the majority of estimated
precipitation from the original SCaMPR occurred in anvil regions (Table 1). Application of the new
rain mask improved the distribution of estimated precipitation from SCaMPR retrievals for all DCS
components.
Spatial analysis for the same 16 days shows that the precipitating area identified by the new rain
mask much more closely resembles the precipitating area identified by radar than the original SCaMPR
or SCaMPR with RH inputs (Table 2). Applying the new rain mask reduces the contribution to the total
areal coverage of estimated precipitation from anvil regions by ~42%. Improvements in estimated
precipitating area for anvil regions are of an order of magnitude, and the high probabilities of detection
for raincore regions are maintained with the new rain mask.
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R
Figure 1. Instantaneous (a) Q2 estimated precipitation rate (mm/hr), (b) GOES-retrieved cloud optical depth, and (c) IR
Temperature (K) at 20:45 UTC April 25 th, 2011. Accumulated (d) Q2 estimated rainfall (areal coverage 33.4%), (e)
estimated rain area (31.1%) from the newly developed algorithm using both cloud optical depth and IR brightness
temperature, and (f) SCaMPR retrieved rainfall (areal coverage 48.3%) over the large domain during 20:00-21:00 25 UTC
April 2011 [Stenz et al. 2014].
Table 1. Percentage of accumulated precipitation that fell in each component of DCSs for 16 days with
widespread convective activity during 2012 over OK. The distribution is given for the ground
observations (MESONET), radar estimates (Q2), the original SCaMPR algorithm (SCaMPR), the
SCaMPR algorithm with relative humidity inputs from the GFS (SCaMPR_RH), the original SCaMPR
algorithm with the new rain mask applied (SCaMPR_Mask), and SCaMPR with RH inputs from the
GFS and the new rain mask applied (SCaMPR_Mask_RH)
Platform
Percentage of
Rainfall in CC
Percentage of
Rainfall in SR
Percentage of
Rainfall in AC
Percentage of
Rainfall in
Unclassified/Thin
Anvil Regions
MESONET
Q2
SCaMPR
SCaMPR RH
SCaMPR_Mask
SCaMPR_Mask_RH
71.01
69.75
12.23
15.46
20.78
21.36
24.55
24.31
30.68
40.48
46.82
48.95
2.46
4.46
35.21
31.15
29.05
26.93
1.97
1.49
21.88
12.91
3.36
2.76
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R
Table 2. Number of pixels identified as precipitating during 16 days with widespread convection over
OK by Q2, SCaMPR, SCaMPR RH, and the newly developed rain mask. The percentages shown
represent the contribution each DCS component makes to the total estimated precipitating area.
Classification
Q2 precipitating
pixels
CC
SR
AC
Error/Thin Anvil
Total
50906
184536
58364
35046
328852
(15.5%)
(56.1%)
(17.7%)
(10.7%)
SCaMPR
precipitating
pixels
40781 (3.8%)
155577 (14.4%)
154345 (14.3%)
731040 (67.6%)
1081743
SCaMPR RH
precipitating
pixels
36363 (4.0%)
131239 (14.4%)
128239 (14.1%)
612861 (67.4%)
908702
New Algorithm
precipitating
pixels
50242 (15.1%)
151638 (45.6%)
41050 (12.3%)
89956 (27.0%)
332886
Figure 2. The bin and total mass distributions as a function of central-bin spectrum (2DC+HVPS, 30-30,000 µm) over the
ice-phase layer for the six selected DCS cases. In each plot, the yellow bar represents the IWC at each individual bin
calculated using refined mass-dimensional relationship, the red-dotted line is the cumulation of yellow bars, representing the
best-estimate total IWC when it reaches the end of spectrum. The blue-dotted line which diverges from the red-dotted line at
900 µm represents the best-estimate total IWC using spectrum less than or equal to 900 µm. The green-dotted line represents
the total IWC directly measured by Nevzorov TWC sensor.
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R
2.3. Dr. Dong’s Ph.D student, Mr. Jingyu Wang, has processed the aircraft in situ measurements
over the ARM SGP site during the MC3E IOP, and submitted the following paper to JGR.
Wang, J., X. Dong, and B. Xi, 2014: Investigation of Microphysical Properties of DCS using
Aircraft in-Situ Measurements: Part I: the Ice-Phase Layer of DCS. Submitted to JGR.
Abstract: The 2011 Midlatitude Continental Convective Clouds Experiment (MC3E) was a highly
successful field campaign with 15 convective cases sampled by the University of North Dakota (UND)
Citation II aircraft in addition to multiple ground-based sensors. There were at least six deep convective
systems (DCS) (25 April, 27 April, 11 May, 20 May, and 23–24 May 2011), including the classic DCS
case on 20 May 2011, which has drawn much attention from observational and modeling communities.
During the experiment, microphysical properties of DCS were studied, including ice water content
(IWC) at upper layer of DCSs. As demonstrated in Fig. 2, the total water content (TWC) measured by
Nevzorov hot-wire sensor could not capture all the ice crystals especially those with sizes greater than
900 µm due to instrumentation limits. This study first eliminated the contamination of super-cooled
liquid water by determining the accurate temperature threshold of pure ice phase layer using multisensors including icing detector records and 2DC image reading results, then recalibrated and developed
a series of new mass-dimensional relationships to generate the best-estimate IWC for each case. Results
indicated that the underestimation ratio of Nevzorov TWC sensor in measuring IWC was not a fixed
factor, but varied depending on the proportion of ice crystals greater than 900 µm in diameter. This
result can be used for the correction of past Nevzorov IWC measurement and retrieval of other
microphysical properties in the pure ice phase layer of DCS.
Fig. 3. (left) The time series of NEXRAD radar reflectivity with corresponding UND aircraft flight (black line) heights and
temperatures. (right) The cloud Particle Size Distributions at different heights and temperatures, measured by UND aircraft
during the MC3E IOP (Wang et al. 2014)
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R
2.4 Dr. Dong’s MS student, Ms. Jingjing Tian has developed a new method to retrieve DCS ice cloud
microphysical properties using ARM cloud radar and aircraft in-situ measurements.
Tian, J, X. Dong, B. Xi, S. Giangrande, and T. Toto, 2014: Retrievals of cloud microphysical properties
of deep convective systems using ARM radar and aircraft in-situ measurements. Part I: ice-phase layer.
Submitted to JGR.
Abstract: This paper presents an algorithm for retrieving cloud microphysical properties in the ice-phase
layer of Deep Convective Systems (DCS) using Atmospheric Radiation Measurement (ARM) Ka-band
cloud radar (KAZR) reflectivity during the Midlatitude Continental Convective Clouds Experiment
(MC3E) at the ARM Southern Great Plain (SGP) site during April-June 2011. It is a challenge to
retrieve the cloud microphysical properties of DCSs due to the attenuation of radar reflectivity, unknown
particle size distributions (PSDs), and the bulk habit of the ice particles within the sample volume. To
address the most pronounced of these radar system limitations, the original KAZR reflectivity
measurements have been adjusted by a collocated unattenuated 915-MHz profiling radar system UHF
ARM Zenith Radar (UAZR) and Joss-Waldvogel impact disdrometer (JWDs). For this study, aircraft insitu measurements provide the best-estimate ice water content (IWC) and particle size distributions
(PSDs) for validating the radar retrievals. With the aid of the scattering database (SCATDB), the
relationships between backscatter cross section () and particle dimension (D) are parameterized for
four ice crystal habits (bullet rosettes, snowflakes, columns, and plates).
The IWC and effective radius (re) in the ice-phase layer of the DCS on 20 May 2011 during the
MC3E have been retrieved from adjusted KAZR reflectivity assuming a modified gamma distribution,
size shape  and a bullet rosette -D relationship. The averaged IWC and re from radar retrievals
over the stratiform rain (SR) region of the DCS are 0.34 g/m3 and 338 µm, in excellent agreement with
aircraft in-situ measured IWC (0.34 g/m3) and re (337 µm). Over the anvil cloud (AC) region, the
retrieved and measured IWCs are 0.18 g/m3 and 0.23 g/m3, and their re values are 250 µm and 305 µm,
respectively. The radar retrieved re and IWC can increase to 283 µm and 0.23 g/m3 if a 2 dBZ
uncertainty is added to the adjusted KAZR reflectivity over the AC region, following the sensitivities of
13%/2 dB in re and 26%/2 dB in IWC.
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R
Figure 4. The 1-min averages of (a) ARM SGP adjusted KAZR reflectivity, (b) radar-retrieved re (black lines) and (c) IWC
(black lines) with corresponding aircraft derived re and IWC values (filled red circles) from 2DC and HVPS measurements at
the same altitudes (~ 7.6 km) as radar retrievals. The grey shaded area represents (a) 2 dB uncertainties of the adjusted KAZR
reflectivity and the range of the retrieved (b) re and (c) IWC in 2 dB uncertainties. The yellow shaded area represents (a) 4 dB
uncertainties of the adjusted KAZR reflectivity and the range of the retrieved (b) re and (c) IWC in 4 dB uncertainties.
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R
2.5. Dr. Dong’s MS student, Ms. Ning Zhou, has collected and processed the hybrid product, Q2
precipitation, and SCaMPR retrievals over the continental USA during the period 2010-2012. This
database will be used to statistically evaluate the SCaMPR retrievals over the different regions of the
continental USA for the future improvement of SCaMPR retrievals.
Figure 5. Upper panel: NEXRAD Q2 product; middle: Hybrid product generated from NEXRAD and GOES data based
on Feng et al. (2011) method; and bottom: SCaMPR precipitation retrievals.
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R
Section B: Dr. Zhanqing Li’s group at University of Maryland at College Park
 Have developed an automatic algorithm to identify DCS using A-train satellite data to obtain
basic statistics of DCS around the globe, which can be used as a benchmark for testing similar
algorithms for GOES-R;
 Have investigated the impact of aerosols on DCS by analyzing more years of matched data
between CloudSat, CALIPSO, MODIS. A past study using one year of data showed that aerosols
invigorate DCS by increasing cloud top height and cloud thickness, rainfall frequency
substantially, but have little impact on low-level liquid clouds. The study will be enriched by
employing much more data samples.
 Have used DCS identified from satellite to assess the performance of NCEP/GFS model in
treating DCS. Our earlier study (Yoo et al. 2012, Zhang et al. 2014) demonstrated that these
products are highly valuable in helping diagnosis model deficiencies. In 2014, we will consider
different types of clouds, especially low stratus clouds and DCSs.
3.1. Dr. Li and his group member, Mr. Jie Peng
3.1a. Using improved automatic algorithm to identify DCS from four years’ A-Train satellite datasets to
obtain the basic statistics of DCS at global scale, which can be used as a benchmark for testing similar
algorithms for GOES-R;
Figure 6 Four year’s trend of the number of DCS (left panel) and cloud top of Core of DCS (right panel) in different zonal
region.
3.1b. They had also been investigating the aerosol impact on DCS by matching four years’ data from
CloudSat, CALIPSO and MODIS following and enriching former study by adding more samples and
active aerosol measurements from CALIPSO, and preparing a paper;
(a)
(b)
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R
(c)
(d)
(e)
(f)
Figure 7. Cloud properties as functions of AI/AOD for (a,b) cloud top temperature and (c,d) cloud
thickness and (e,f) cloud ice water path over land (left panles) and ocean (right panels). The right-hand
axes of (a,b) are for the liquid clouds
Abstract: Using a large ensemble of satellite data from CloudSat, the Cloud-Aerosol Lidar and Infrared
Pathfinder Satellite Observations platform and the moderate resolution imaging spectroradiometer
onboard the Aqua platform over the tropical region of the world, we show how there is a significant
influence of cloud properties with increasing aerosol index(AI) over oceans and aerosol optical
depth(AOD) over land for mixed-phase clouds; no significant change was seen for liquid clouds. Further
investigate indicates that the observed phenomenon can be explained via microphysical and invigoration
effects. Considering the inherent limitation of polar orbit satellite observations, the causal relationships
between cloud properties and AI/AOD cannot be fully supported by our findings. But if the observed
phenomena are indeed due to the effect of aerosols, this would have a strong influence on Earth’s
radiation budget, thus leading to a long-term impact on climate change.
3.1c Dr. Li’s former Ph. D student, Dr. Hyelim Yoo published a paper entitled: “Evaluation of cloud
properties in the NOAA/NCEP global forecast system using multiple satellite products” and submitted a
paper to Climate Dynamics: “Diagnosis and testing of low-level cloud parameterizations for the
NCEP/GFS model using satellite and ground-based measurements”
Abstract: Clouds are one of the most critical factors in Earth’s climate system and they play an
important role in regulating Earth’s radiative energy budget (Barker et al., 1999; Collins 2001). Cloud
representations in models are a critical source of uncertainty in predictions of climate change. However,
many GCMs have difficulties in representing low cloud distributions, especially over eastern tropical
oceans. This is also a problem for fully-coupled climate system models and regional climate models.
This inability of climate models to simulate marine stratocumulus cloud decks is one of the most
significant sources of error and uncertainty associated with cloud feedback processes (IPCC, 2007).
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R
Using datasets capable of detecting multi-layer clouds and their optical properties, we can
evaluate cloud properties forecasted by the National Centers for Environmental Prediction (NCEP) GFS
model. Over all, the GFS-generated cloud properties such as cloud fraction, cloud optical depth, liquid
and ice water paths, and the frequencies of multi-layer clouds match with satellite retrievals reasonably
well in terms of locations and spatial patterns, although large discrepancies exist in their magnitudes
(Yoo and Li 2012). The most outstanding problem lies in a severe underestimation of low clouds over
oceans by the GFS model. To find out the causes for the discrepancies, this study conducts analysis of
input variables needed by cloud schemes and examination of cloud parameterization schemes. Simulated
temperature and relative humidity profiles are examined against retrievals from the Atmospheric
InfraRed Sounder (AIRS) sensor and ground-based measurements. The evaluation of atmospheric
environmental variables allows investigation of whether, and how much, discrepancies in low cloud
distributions between the model and observations may result from model input parameters or from
deficiencies in model cloud parameterization schemes. An alternative cloud parameterization scheme
based on a diagnostic approach is applied to the GFS model in order to improve the simulations of
clouds by the GFS model.
Figure 8. Low cloud fraction from the CL algorithm (left plot) and the GFS model (right plot) during
July 2008.
Figure 8 compares the global distributions of low-level cloud fraction estimated from the CL algorithm
applied to MODIS data and the GFS model simulations. Simulated low clouds agree with satellite
retrievals in terms of the location of cloud layers. While the satellite retrievals show extensive marine
stratocumulus clouds over the eastern tropical Pacific and Atlantic oceans, such clouds are not well
simulated in the GFS model. This led to the overestimation (underestimation) of outgoing longwave
(shortwave) fluxes in the GFS model at the top-of-atmosphere, emphasizing the impact of marine
stratocumulus clouds on global net radiation. The GFS model produces less upwelling SW radiation and
more outgoing LW radiation than that measured by CERES and these radiation errors are co-located
with the regions of under-estimated stratocumulus cloud cover.
3.1d. Dr. Li’s team has exploited a variety of satellite products to identify deep convective cloud system
(DCC), analyzed its spatial and temporal variations for understanding their impact on precipitation and
radiation budget.
Clouds and their interactions with atmospheric circulation, shortwave and longwave radiation and
the surface are very important components of both weather and climate. Serving as one of the most
important elements of hydrological and energy circulation, DCS have a crucial impact on not only
radiation budgets at the surface and the top-of-the-atmosphere, but also on heating profiles within the
atmosphere and the spatial and temporal distribution of precipitation around the globe. Usually
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R
accompanied by precipitation, DCS have a more direct impact on hydrologic circulation. Using four
years’ worth of CloudSat and CALIPSO data (2007-2010), DCS were identified and the global
distribution of the number of DCS, TDCC, DDCC and WDCS was analyzed. The main conclusions are
as follows:
(1) The greatest concentration of DCS occurs in central Africa, northern South America, northern
Australia and Tibet; maximum values of NDCS in a single grid box within each of these regions over
the four-year period are 86, 112, 87 and 154, respectively. There is a significant increase in the NDCS
over the Antarctic polar region during the NH summer/SH winter. In the NH, the least NDCS are seen in
winter (9,570) and the greatest NDCS occurs in summer (25,341); the total number of DCS in spring
and autumn are 15,141 and 20,021. There is more seasonal and zonal variation in NDCS in the SH. The
maximum NDCS seen in the LL occurs in summer (9479) and the minimum, in winter (4222); at HL in
the SH, the reverse is seen (winter: 5705, summer: 801). The NDCS remains generally the same yearround in the ML of the SH.
(2) TDCC decreases towards higher latitudes in both hemispheres. DCS with the highest TDCC (up
to 16 km) and largest DDCC (~15 km) occur over south and eastern Asian monsoon regions, westcentral Africa and northern South America. There is little seasonal variation in mean TDCC in the LL
and HL of the NH. The mean TDCC in the ML of the NH reaches a maximum in summer (10.59km)
and a minimum in winter (9.89km).In the SH, the highest TDCC occurs in the LL and ML in summer
(13.51km and 11.03km, respectively) and the lowest TDCC occurs in these regions in winter (13.00km
and 10.77km). The mean TDCC in HL of the SH vary greatly from month-to-month but little difference
in TDCC between different seasons is seen.
(3) The WDCS increases in magnitude towards higher latitudes; mean WDCS in the LL, ML and
HL of both hemispheres is 665.05km, 1066.44km and 1218.85km, respectively. Over the four-year
period, DCS with the largest WDCS were found in the HL of the SH. The maximum and minimum
WDCS in the NH is 1383.1km (winter) and 645.2km (summer), respectively. The WDCS in the LL of
the SH is largest in summer (802.8 km) and smallest in winter (585.7 km).In the ML of the SH, the
maximum (minimum) WDCS (1337.5 km; 1021.1 km) occurs in winter (summer). The largest and
smallest WDCS are seen in autumn and winter in the HL of the SH (1364.7 km and 1168.8 km,
respectively).
(4) Seasonal variations of TDCC and DDCC in different zonal regions are mainly due to the
different mechanisms for generating DCS at different latitudes. DCS at LL are mainly deep convective
clouds, highly developed in the vertical direction but covering a relatively small area in the horizontal
direction. Most DCS in the ML are formed by frontal systems so have smaller TDCC than those in the
LL. The NDCS in the ML of the NH (SH) increases significantly in summer (winter) because of the
monsoon circulation. The DCS at HL are mainly generated by large frontal systems so have the largest
WDCS and smallest TDCC.
Future work will focus on examining the AIE in DCS at a global scale.
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R
Figure 9. Example of a DCS identified on February 4, 2007.
Figure 10. Total number of DCS in each 5ox5o grid box over the four-year period for (a) NH spring/SH
autumn, (b) NH summer/SH winter, (c) NH autumn/SH spring, and (d) NH winter/SH summer.
White grid boxes signify that no DCS were identified in that box.
3.1c Dr. Li’s team published a paper entitled: “Cloud Vertical Distribution from
Radiosonde, Remote Sensing, and Model Simulations” Climate Dynamics, in press.
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R
Knowledge of cloud vertical structure is important for meteorological and climate studies due to the
impact of clouds on both the Earth’s radiation budget and atmospheric adiabatic heating. Yet it is
among the most difficult quantities to observe. In this study, we develop a long-term (10 years)
radiosonde-based cloud profile product over the Southern Great Plains (SGP) and along with groundbased and space-borne remote sensing products, use it to evaluate cloud layer distributions simulated
by the National Centers for Environmental Prediction (NCEP) Global Forecast System (GFS) model.
The primary objective of this study is to identify advantages and limitations associated with different
cloud layer detection methods and model simulations. Cloud occurrence frequencies are evaluated on
monthly, annual, and seasonal scales. Cloud vertical distributions from all datasets are bimodal with a
lower peak located in the boundary layer and an upper peak located in the high troposphere. In general,
radiosonde low-level cloud retrievals bear close resemblance to the ground-based remote sensing
product in terms of their variability and gross spatial patterns. The ground-based remote sensing
approach tends to underestimate high clouds relative to the radiosonde-based estimation and satellite
products which tend to underestimate low clouds. As such, caution must be exercised to use any single
product. Overall, the GFS model simulates less low-level and more high-level clouds than observations.
In terms of total cloud cover, GFS model simulations agree fairly well with the ground-based remote
sensing product. A large wet bias is revealed in GFS-simulated relative humidity fields at high levels
in the atmosphere.
Fig. 11 Annual mean cloud occurrence frequencies from 2001 to 2010 for all-clouds (black line), low- (pink line), mid(green line), high-level clouds (blue line),and deep clouds (red line) derived from (a) radiosonde data, (b) the ARSCL
product, and (c) the GFS model at the SGP site.
Additional Information
1. Interaction with operational partners –
We have closely collaborated with Dr Bob Kuligowski, NOAA NESDIS to improve the current
SCaMPR precipitation retrievals as reflected in our paper(s). So far, we have provided him with
our rain mask algorithm so it can be incorporated into the operational version of SCaMPR.
Additionally, we have evaluated multiple versions of the SCaMPR algorithm for Dr. Kuligowski
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R
including a recent version incorporating relative humidity (RH) inputs from the GFS model to
identify strengths, weaknesses, and biases.
2. Conference/workshop participation –
Mr. Ron Stenz submitted two posters NOAA GOES-R meetings in 2013 and 2014, and one
poster for 2013 Gordon conference meeting; Jingyu Wang and Jingjing Tian, have attended both
spring and fall DOE ASR Science Team meetings in 2013 and 2014 and presented their results
during the meetings. Special sections on cloud remote sensing were convened by the PI and coPI at the AGU Fall Meeting in 2013, and the Asian Atmospheric and Oceaninc Conferences in
Sapporo Japan, 2014.
3. Outside project publicity –
The PI group generated hybrid classification product and collected NEXRAD Q2 product have
drawn a great interest from DOE ARM and NOAA MAPP communities, and we have provided
these products to more than 10 modelers to improve their model simulations and reanalysed
results.
4. Plans for operational transition –
We have already provided Dr. Kuligowski with the algorithm for the rain mask to improve the
performance of the SCaMPR algorithm. Currently, preliminary SCaMPR products are made
available to the public at www.star.nesdis.gov/smcd/emb/ff/SCaMPR.php. Updates and
improvents to SCaMPR will be made during continued collaboration with Dr. Kuligowski, as the
product is refined to meet the needs of forecasters and users.
5. Journal articles –
The following papers from Dr. Xiquan Dong’s group
a) Feng, Z., X. Dong, B. Xi, S. McFarlane, A. Kennedy, B. Lin, and P. Minnis, 2012: Life cycle of deep
convective systems in a Lagrangian Framework. JGR. 117, D23201.
b) Wu, D., X. Dong, B. Xi, Z. Feng, A. Kennedy, G. Mullendore, M. Gilmore, and W-K Tao, 2013,
The Impact of Various WRF Single-Moment Microphysics Parameterizations on Squall Line
Precipitation, JGR, 118, 1-17, DOI: 10.1002/jgrd.50798.
c) Stenz, R., X. Dong, B. Xi, and B. Kuligowski, 2014: Assessment of SCaMPR and NEXRAD Q2
precipitation estimates using Oklahoma Mesonet observations. J. Hydrometeorology (In press).
d) Tian, J., X. Dong, B. Xi, S. Giangrande, and T. Toto, 2014: Retrievals of cloud microphysical
properties of DCS using ARM ground-based and aircraft in situ measurements. Part I: Ice-phase
layer. Submitted to JGR.
e) Wang, J., X. Dong, and B. Xi, 2014: Investigation of Microphysical Properties of DCS using
Aircraft in-Situ Measurements. Part I: the Ice-Phase Layer of DCS. Submitted to JGR.
f) Lin, C., Z. Pu, X. Dong, and S. Krueger, 2014: Evaluation of Double-Moment Representation of
Warm-Rain and Ice-Rain Hydrometeors in Bulk Microphysics Parameterization. To be submitted to
Mon. Wea. Rev.
g) Fan, J., Y-C. Liu, K. Xu, G. Zhang, X. Dong, K. North, S. Collis, C. Qian, and S.Ghan, 2014:
Development of a Scale-Aware Cumulus Parameterization – Part I: Evaluation of Model
Simulations with Bin Microphysics and Comparisons with Bulk Microphysics. Submitted to
JGR.
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R
The following papers from Dr. Zhanqing Li’s group
h) Niu, F., and Z. Li, 2012: Systematic variations of cloud top temperature and precipitation rate with
aerosols over the global tropics, Atmos. Chem. Phys., 12, 8491-8498.
i) Yoo, H., and Z. Li, 2012: Evaluation of cloud properties in the NOAA/NCEP global forecast system
using multiple satellite products, Clim. Dyn., doi:10.1007/s00382-012-1430-0.
j) Fan, J., D. Rosenfeld, Y. Ding, L. R. Leung, and Z. Li, 2012: Potential aerosol indirect effects on
atmospheric circulation and radiative forcing through deep convection, Geophys. Res. Lett., 39,
L09806, doi:10.1029/2012GL051851.
k) Fan, J., D. Rosenfeld, Y. Ding, L. R. Leung, and Z. Li, 2012: Potential aerosol indirect effects on
atmospheric circulation and radiative forcing through deep convection, Geophys. Res. Lett., 39,
L09806, doi:10.1029/2012GL051851.
k) Zhang, J., Z. Li, H. Chen, M. Cribb, H. Yoo, 2014, Evaluation of Cloud Structure Simulated by NECP
GFS using Ground-based Remote Sensing and Radiosonde Products at the ARM SGP site, Climate
Dynamics, in press.
l) Yang X and Z Li. 2014. "Increases in thunderstorm activity and relationships with air pollution in
southeast China." Journal of Geophysical Research – Atmospheres, 119(4),
doi:10.1002/2013JD021224.
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R
Key Graphics
Figure 112. Probability Density Functions (PDF) of rain area over entire OK state using a bin width of
10 % during the Midlatitude Continental Convective Clouds Experiment (MC3E) campaign (14 days
with convection). The 0.25 mm threshold was used for both Q2 and SCaMPR to determine whether or
not a pixel was classified as raining.
Conclusion: The newly developed algorithm identified precipitation coverage is almost the same as
that derived from NEXRAD Q2 product, while SCaMPR is overestimated.
07/27/2016
GOES-R3 Status Report Template
NESDIS STAR GOES-R