VISIT (Level 2) Quality Control Report for COARE Surface Meteorological Data

J. Parks Camp and Shawn Smith

Tropical Ocean Global Atmosphere (TOGA)
Coupled Ocean Atmosphere Response Experiment (COARE)

Surface Meteorological Data Processing Center
Center for Ocean Atmospheric Prediction Studies
The Florida State University

July 31, 1996

Report COAREMET 96-4
Version 1.0

1. Introduction
The Data Processing Center (DPC) for Surface Meteorology at Florida State University (FSU) was charged with collecting, quality controlling, archiving, and distributing surface meteorological data from research vessels, buoys, and land stations during the Tropical Ocean Global Atmosphere (TOGA) Coupled Ocean Atmosphere Response Experiment (COARE). The DPC collected multiple surface meteorological records from the COARE intensive observing period (IOP; 01 November 1992 through 28 February 1993). Types of data collected include standard ship bridge observations, advanced automated buoy, shipboard, and land systems; e.g. Improved METeorology (IMET) measurements system, Integrated Sounding System (ISS), Coastal Marine Automated Network (CMAN), along with all other practically obtainable surface meteorological data from COARE research vessels, land stations, and buoys.

The overall goal of the DPC's quality control (QC) is to provide a well-documented, reliable, and consistent data set to the scientific community. The DPC evaluated the data to the highest possible level with the QC results depending primarily on the level of knowledge of the incoming data as well as examination and scientific analysis of the data. Data quality flagging, not value replacement, was the primary method used to denote suspect or erroneous data. The choice to flag questionable data at the DPC, as opposed to data correction, was made to insure the data remain in the original form that was sent to the DPC. For scientific reasons, the end user must determine appropriate utilization of the flags provided.

The QC procedure for all COARE data sets began with the acquisition of a COARE meteorological data set and accompanying metadata (e.g. README file). The original data arrived at the DPC in varying formats from digital to hard copy logs. Each data set contained different meteorological variables and the data were collected at varying time intervals. As a result, the first step of the QC was converting the meteorological data to a standard format. The format includes metadata and is described in detail in Smith and Legler (1995a, 1995b).
Once converted, the data underwent a three step QC process beginning with an automated preprocessor that checked for values outside of a realistic range, statistical outliers, unrealistic ship movements, etc. During the second QC stage, the preprocessed meteorological data were visually inspected by a trained Data Quality Evaluator (DQE) using the VIsual Data Assessment Tool (VIDAT). VIDAT is an interactive, graphically-based tool developed at the FSU DPC, that allows the DQE to add or remove QC flags from a time series of meteorological data. In the third step of the QC, the DQE compared data at neighboring stations using the VIsual Station Intercomparison Tool (VISIT). VISIT is another interactive, graphically-based tool developed at FSU. Details of the three stage QC process are found in Smith et al. (1996).

The focus of this report is to outline the problems and interesting features discovered in the COARE surface meteorological data during the VISIT stage of the QC process. Section 2 outlines the platforms and meteorological values evaluated with VISIT. A discussion of problems found, flagged, or corrected for each meteorological variable is found in section 3. Section 4 contains a summary statement of data quality and flag statistics for each of the platforms. These summaries are the final result of all three stages of the QC process.

2. Data Evaluated
VISIT processing was initiated when all available COARE surface meteorological data had completed VIDAT testing. Table 1 contains a list of the platforms included in VISIT for COARE. Platforms not evaluated in VISIT processing include R/V's Alis, Kaiyo, and Keifu Maru (Data were never received from these R/V's) and data from the Priority Sounding Stations (PSS). The PSS data were received too late for inclusion.

Table 1: List of independent meteorological data sets reviewed within the VISIT analysis for COARE.

Station Mounting Platform Name/Type Number of Independent Data Sets

Moored Buoy ATLAS 15

PROTEUS 2

VAWR/IMET 1

Drifting Buoy Marisonde 4

Niiler 83

Land CMAN 7

GTS 79

ISS 4

Research Vessel

Franklin
AWS, Bridge 2

Hakuho Maru
Bridge 1

Kexue #1
ISS 1

Le Noroit
AWS 1

Malaita
Bridge 1

Moana Wave
AWS 1

Natsushima
Bridge 1

Shiyan #3
ISS, Bridge 2

Vickers
AWS 1

Wecoma
AWS 1

Xiang Yang Hong #5
Bridge 1

Total 208

Note: No surface meteorological data for the R/V Alis, Kaiyo, and Keifu Maru were ever made available to the DPC. Also, due to multiple data problems, the PSS data were not reviewed in VISIT.

Using VISIT, the data were displayed visually to allow the DQE to look for anomalous values, unrealistic gradients, or physically inconsistent data. When questionable values are located, the DQE next determined which flag to apply to the value. Flags available to the DQE are listed in Table 2 and described in detail in Smith et. al (1996).

Table 2: Definitions of COARE quality control flags and reference page numbers.

Flag Definition

A Original data had unknown units. The units shown were determined using a climatology or some other method.

B Original data were out of the range bounds outlined.

C Time data are not sequential or date/time not valid.

D Data failed T„Tw„Td test. In the free atmosphere, the temperature is always greater than or equal to the wet-bulb temperature, which in turn is always greater than or equal to the dew point temperature.

E Data failed resultant wind recomputation check. When the data set includes all variables required, a program recomputes the earth relative wind speed and direction and compares the computed values to the reported earth relative wind speed and direction. A failed test occurs when the wind direction difference is > 10š or the wind speed difference is > 5 m/s.

F Platform velocity unrealistic. Determined with platform position data.

G Data are > 4 standard deviations from the COADS climatological means (da Silva et al. 1994). Test applied only to pressure, temperature, sea temperature, relative humidity, and wind speed.

H Discontinuity found in data.

I Interesting feature found in data. More specific information on the feature is contained in the QC reports. Examples include: hurricanes, sharp sea water temperature gradients, strong convective events, etc.

J Data are of poor quality by visual inspection, DO NOT USE.

K Data suspect - USE WITH CAUTION - flag applied when the data look to have obvious errors, but no specific reason for the error can be determined.

L Oceanographic platform passes over land.

M Known instrument malfunction.

N Inconsistent with neighboring station.

P Position of platform or its movement are uncertain. Data should be used with caution.

Q Data arrived at the DPC already flagged as questionable.

S Spike in the data. Usually one or two sequential data values (sometimes up to four values) that are drastically out of the current data trend. Spikes occur for many reasons, including power surges, typographical errors, data logging problems, lightning strikes, etc.

T Time duplicate.

V Colocated data differ - use with caution.

Z Data passed evaluation.

Table 1: List of independent meteorological data sets reviewed within the VISIT analysis for COARE.
Station Mounting	Platform Name/Type	Number of Independent Data Sets
Moored Buoy	ATLAS	15
PROTEUS	2
VAWR/IMET	1
Drifting Buoy	Marisonde	4
Niiler	83
Land	CMAN	7
GTS	79
ISS	4
Research Vessel
Franklin	AWS, Bridge	2
Hakuho Maru	Bridge	1
Kexue #1	ISS	1
Le Noroit	AWS	1
Malaita	Bridge	1
Moana Wave	AWS	1
Natsushima	Bridge	1
Shiyan #3	ISS, Bridge	2
Vickers	AWS	1
Wecoma	AWS	1
Xiang Yang Hong #5	Bridge	1
Total	208
Note: No surface meteorological data for the R/V Alis, Kaiyo, and Keifu Maru were ever made available to the DPC. Also, due to multiple data problems, the PSS data were not reviewed in VISIT.

Hourly data from the entire IOP were analyzed one variable at a time. Higher temporal resolution data were flagged when necessary. Pressure (P), temperature (T), dewpoint (TD), sea temperature (TS), wet-bulb temperature (TW), relative humidity (RH), earth relative wind speed (SPD), and earth relative wind direction (DIR) were quality controlled.

3(a). Pressure
Overall, the intercomparison of pressure data from the TOGA COARE IOP showed the pressure values to be in excellent condition. Aside from sensor malfunctions and other minor problems found during VIDAT processing, most of the data values were consistent with values from nearby reporting stations and platforms. The data from several stations, however, were noticeably problematic.

The atmospheric pressure values reported by Lae (WMO ID #94027) were very inconsistent with surrounding observing platforms. Less than ten values for pressure were reported during the IOP and of these, nearly all of them differed dramatically from nearby observations. The values from Lae were typically near 1025 mb while values at the nearby stations of Wewak (94004), Manus (94044), and Rabaul (94085) were below 1010 mb. The data were flagged with 'N' flags and these values should be used with extreme caution.

Atmospheric pressure data from the Vickers standard observations, Shiyan ISS, and Kexue ISS were noticed to be consistently lower than values from surrounding platforms. The difference in the pressures from surrounding observations for each ship was approximately 2 mb. After some investigation, it was concluded that the differences could be attributed to the height of the pressure sensors aboard each ship. It was confirmed that the height of the pressure sensor for the Kexue ISS data was 24 m which would account for the 2 mb difference between this and the sea level pressure. The height of the sensors aboard the Shiyan and Vickers were unavailable but a similar scenario is likely.

Data obtained from Kapingamarangi between December 9, 1992 and December 14, 1992 was extremely noisy when compared to nearby ISS stations. The values varied by several millibars over the space of a few minutes (see Figure 1). The general trends, however, were in agreement, and no explanation for the noise was available. These values were not flagged but the data from these days should be used with some caution.

Figure 1: Atmospheric Pressure data from Kapingamarangi for December 12-13, 1992. Note the noisy condition of the data during the period.

Though the pressure data from the research vessel Franklin were in good condition, a time shift was evident in the data. Upon inspection, the data values appeared to be shifted by eight to ten hours from values being reported at nearby stations. Upon arrival at the DPC, the time stamp was thought to be UTC. However, after the time shift was noticed, the DPC confirmed that the Franklin data arrived with the times given in Eastern Australian Standard Time which is ten hours ahead of UTC. A similar problem was discovered in the data from the Xiangyanghong. Like the Franklin, these data appeared to be shifted by eight or ten hours. However, we were not able to confirm whether or not the data had been recorded using local time. As a result, ten hours were subtracted from all time stamps for the Franklin and Xiangyanghong bridge data. The resulting time series compared well with neighboring stations upon a second VISIT inspection.

Other than these large scale problems, there were some individual pressure values that were inconsistent with surrounding observations.

3(b). Air Temperature
The majority of the air temperature data from the COARE IOP are in excellent condition. Most of the data values are consistent with values from nearby reporting stations and platforms. Temperature was one of the most difficult variables to quality control because of the different platforms that were used to collect this data. Each platform (R/V, buoy, and land) resides in an environment which reacts differently to sunrise, sunset, clouds, etc. With this in mind, only the values that deviated greatly from their surroundings were flagged.
As with pressure, the temperature values reported by Lae (94027) were very inconsistent with surrounding observing platforms. Less than ten values for temperature were reported and of these, most were between twenty and forty degrees below observations from nearby platforms. These data were flagged appropriately and should not be used.
Noise problems similar to those that appeared in the Kapingamarangi pressure data appeared in temperature data from the Shiyan ISS (BXMC) and Kexue ISS (BCCU). The temperature values varied by as much as ten degrees over a few minutes to an hour while nearby temperature observations were steady (Figure 2). The most significant of the temperature fluctuations on each of these ships were flagged as erroneous or the standard deviation flag from the prescreener was left unchanged.

Figure 2: Air temperature data from (top) Kexue and (bottom) Atlas buoy 52012 (-2.0š N, 156.0š E) for December 17, 1992. Note noisy condition of Kexue data.

A sporadic conversion problem was discovered in the GTS temperature data. During VIDAT processing, numerous temperature values in the GTS data were flagged as spikes. In VISIT, however, it was discovered that the spikes could be the correct temperatures but had been encoded in Fahrenheit instead of Celcius. Therefore, the Fahrenheit temperatures were converted to Celcius and the flags were removed (Table 3). The modified values were reviewed; many of the corrected values fit in well. However, several corrected values were unusually high or low and remained flagged as spikes.

Table 3: Number of šF to šC conversions by station

Station Name Station ID Number of šF to šC Conversions

Lukunor Atoll 94339 2

Ponape Island 91348 1

Jaluit Atoll/Jabor 91369 1

Majuro/Marshall Island 91376 1

Funafuti Intl Airport 91643 1

Padang/Tabing 96163 2

Biak/Mokmer 97560 1

TOTAL 9

Other than these problems, the rest of the temperature data was in excellent condition.

3(c). Wind
Only the wind data from research vessels were quality controlled using VISIT because direct intercomparison between land, ship, and buoy winds cannot be done accurately without correcting all winds to a common height reference. Wind sensors on buoys are closer to the surface than ship wind sensors, and land stations are at various elevations. The wind speeds and directions were plotted as vectors to make the quality control process easier. Aside from values flagged in earlier QC stages, nearly all of the wind data from the research vessels during the TOGA COARE IOP are in excellent condition. No flags were added to these data during VISIT processing. However, a problem with the wind data from the R/V Franklin (VJJF) was indicated.

Two separate wind data sets were obtained from the Franklin. The research quality data included wind direction and speed from one sensor at 11.8 meters. The bridge observations from the Franklin included wind direction and speed from two different sensors, one at 10 meters and the other at 18 meters. During VISIT processing, it was observed that the research and bridge wind direction data differed dramatically. The difference between the directions appeared to be close to or exactly 180 degrees (Figure 3).

(a)                                                           (b)

Figure 3: Histograms of differences in wind direction between (a) 11.8 m and 10 m sensors, and between (b) 11.8 m and 18 m sensors.

The wind directions for either the bridge or research data are incorrect by 180 degrees. Using VISIT, it was determined that the Franklin research quality wind direction data agreed with neighboring stations more so than the bridge data. Therefore, 180š were added to the wind directions from the bridge observations (at both the 10 m and 18 m heights). Revised difference histograms that include the modified bridge winds clearly show an improvement. The scatter shown in Figure 4b is most likely a function of recording winds at different sensor heights.

(a)                                                           (b)

Figure 4: Histograms of differences in wind direction between (a) 11.8 m and corrected 10 m sensor data, and between (b) 11.8 m and corrected 18 m sensor data.

Other than the direction problem with the Franklin, no other problems were found in the wind data from the research vessels.

3(d). Dewpoint
Aside from the problems found in VIDAT processing, the dewpoint data from the TOGA COARE IOP are in excellent condition. Nearly all of the data values are consistent with values from nearby reporting stations and platforms. Only a few "N" flags (Table 2) were added.

3(e). Sea Temperature
Aside from minor problems found in VIDAT processing, the sea temperature data from the TOGA COARE IOP are in excellent condition. Very few flags were added to this data as most of the sea temperatures were in agreement with data values from nearby reporting stations and platforms.

3(f). Relative Humidity
Aside from the major problems found in VIDAT processing at Kapingamarangi, the remaining relative humidity data from the TOGA COARE IOP are in good condition. Most of the data values are consistent with values from nearby reporting stations and platforms. Only a few minor discrepancies were found in the data and these were flagged appropriately.

4. FLAG SUMMARIES

*Buoys--WHOI*
Variable	Out of Bounds	Erroneous Data	Total Number of Flags	Percent of Variable Records Flagged
RH	2		2	0.06
RR		228	228	7.13
Total	2	228	230
Percent of Total Data Flagged	0.0063	0.7134	0.72

Most of the data from the WHOI buoy are considered to be in good condition. The only major problem was found in the rain rate (RR) data where an average value of 0.6 mm/hr were used by WHOI to fill in gaps in the data. These data were flagged as erroneous and should not be used.

*Buoys--PMEL*
Variable	> 4 S.D. From Climatological Mean	Suspect Data	Spike	Total Number of Flags	Percent of Variable Records Flagged
DIR		171		171	0.39
SPD	10	171		181	0.41
T	1			1	0.00*
TS		50	5	55	0.13
Total	11	392	5	408
Percent of Total Data Flagged	0.0031	0.1120	0.0014	0.12

*Less than 0.005% of this variable flagged.

The PMEL data are considered to be in good condition. The only significant problem occurred in the wind direction and wind speed of buoy 52011 and buoy 52012. Flags were added to these buoys when it was noticed that, on several different occasions the wind speed and direction remained constant for several hours at a time. Another minor problem occurred on buoy 52011 when it was noticed that on two occasions, the sea temperature data became somewhat erratic. These data were flagged as suspect.

*Drifter--Marisonde*
Variable	Unrealistic Platform Movement	Spike	Total Number of Flags	Percent of Variable Record Flagged
LAT	4		4	0.09
LON	4		4	0.09
SPD		1	1	0.02
Total	8	1	9
Percent of Total Data Flagged	0.0244	0.0031	0.03

The data from the Marisonde GT drifting buoys are considered to be in excellent condition. Only a few minor problems resulted in nine flags being added.

Drifter--Niller

The Niller drifter buoy data are considered to be in excellent condition. No flags were added to this data.

*Land--CMAN*
Variable	Interesting Feature	Erroneous Data	Spike	Colocation Error	Total Number of Flags	Percent of Variable Records Flagged
DIR		2827	15	37	2879	14.68
SPD		2827	16	35	2879	14.68
P	1			2857	2858	14.57
T	1			2860	2928	14.93
Total	2	5654	99	5789	11544
Percent of Total Data Flagged	0.0015	4.1183	0.0721	4.2166	8.41

Much of the CMAN data are considered to be in good condition, however many of the wind speed and direction values were flagged as bad data. These flags were applied to the data from Mili (91377) which reported an unlikely constant value of 0.0 m/s continuous over a four month period. Also, many colocation flags were added. These flags were added when CMAN data did not agree with a collocated GTS data set. These data should be used with caution.

*Land--GTS*
Variable	Out of Bounds	Interesting Feature	Suspect Data	Intercom parison Problem	Spike	Colocati on Error	Total Number of Flags	Percent of Variable Records Flagged
DIR					2	50	52	0.12
SPD		4			33	24	61	0.14
P		15	1	21	60	969	1066	2.39
T	2		1	20	65	952	1040	2.33
TD	8			9	39		56	0.13
Total	10	19	2	50	199	1995	2275
Percent of Total Data Flagged	0.0028	0.0053	0.0006	0.0140	0.0557	0.5581	0.64

Most of the GTS land station data are in good condition. Spikes were fairly common and often were present in each variable at a particular time. Some spikes were possibly caused by typographical errors by the persons entering the data. A few intercomparison problems were also found in this data set, and many colocation flags were added. The data with colocation flags should be used with caution.

*Land--Kapingamarangi ISS*
Variable	Out of Bounds	>4S.D. From Climatological Mean	Interesting Feature	Erroneous Data	Suspect Data	Instrument Malfunction	Inter- com parison Error	Spike	Total Number of Flags	Percent of Variable Records Flagged
SPD					15				15	0.01
P	1440	35			478			444	2397	1.40
T	61		2	17	1595	966	343	2	2986	1.74
RH					145	103626	235	5	104011	60.78
Total	1501	35	2	17	2233	104592	578	451	109409
Percent of Total Data Flagged	0.0975	0.0023	0.0001	0.0011	0.1450	6.7910	0.0375	0.0293	7.10

Some of the data from the Kapingamarangi ISS Station is considered to be in poor condition. During the period from December 10, 1992 to December 14, 1992, the pressure data is very erratic. Many of these values are flagged as out of realistic range or as being greater than four standard deviations from the da Silva climatological mean (da Silva, et al. 1994). Rhythmic spikes also occurred on several occasions in the pressure data. A malfunction in the humidity sensor caused much of the relative humidity data to be flagged, as well as some of the temperature data since the sensors were coupled. The data analysts suggest caution be used with the pressure, temperature, and relative humidity data from Kapingamarangi.

*Land--Kavieng ISS*
Variable	Interesting Feature	Instrument Malfunction	Total Number of Flags	Percent of Variable Records Flagged
DIR	1		1	0.00*
SPD	1		1	0.00*
P		5728	5728	3.74
T	10		10	0.01
RH	10		10	0.01
Total	22	5728	5750
Percent of Total Data Flagged	0.0016	0.4151	0.42

*Less than 0.005% of variable flagged

The data from the Kavieng ISS land station are considered to be in good condition. The only major problem occurred between 0000 UTC on November 1, 1992 and 2400 UTC on November 15, 1992, when a malfunctioning pressure sensor reported values of about 800 mb.

*Land--Manus ISS*
Variable	Out of Bounds	Interesting Feature	Total Number of Flags	Percent of Variable Records Flagged
DIR		2	2	0.00*
SPD		11	11	0.01
P		1	1	0.00*
T		22	22	0.01
RH	10440	12	10452	6.47
Total	10440	48
Percent of Total Data Flagged	0.7179	0.0033	10488	0.72

The data from the Manus ISS Land station are considered to be in good condition. The only problem occurs in the relative humidity data. The flags added to the relative humidity are due to values being greater than 100%.

*Land--Nauru ISS*
Variable	> 4 S.D. From Climatological Mean	Suspect Data	Spike	Total Number Flags	Percent of Variable Records Flagged
SPD			1	1	0.00*
P	1467	7	381	1855	1.10
RH			1	1	0.00*
Total	1467	7	383	1857
Percent of Total Data Flagged	0.0967	0.0005	0.0252	0.12

*Less than 0.005% of variable flagged

The data from the Nauru ISS Land station are considered to be in good condition. The only major problems occurred in the pressure data, where rhythmic spikes occurred frequently. Also, a large number of pressure values were flagged as being greater than four standard deviations from the da Silva climatological mean (da Silva, et al. 1994) and should be used with caution.

Research Vessel--Natsushima (7JDU)

The data from the research vessel Natsushima are considered to be in excellent condition. No flags were added to this data set.

Research Vessel--Xiangyanghong (BNTA)
Variable	Units Added	Failed T„Tw„Td Test	Unrealistic Platform Movement	> 4 S.D. From Climatological Mean	Intercom parison Error	Time Duplicate	Total Number of Flags	Percent of Variable Records Flagged
TIME	599					2	601	100.00
LAT	596		5				601	100.00
LON	596		5				601	100.00
DIR	601						601	100.00
SPD	597			4			601	100.00
P	601						601	100.00
TD	600	1					601	100.00
RH	601						601	100.00
T	600	1					601	100.00
TW	600	1					601	100.00
TS	600				1		601	100.00
Total	6591	3	10	4	1	2	6611
Percent of Total Data Flagged	99.6975	0.0454	0.1513	0.0605	0.0151	0.0303	100.00

Most of the data from the research vessel Xiangyanghong is considered to be in good condition. A few major problems did occur however. The data arrived at the DPC without any indication as to what units were used for the variables. Thus, units were determined by comparing values on the Xiangyanghong with nearby vessels. Through inspection, it was determined that the time stamps on this data set were in local time and not UTC. Ten hours were added to the time stamps, however, as no confirmation of this correction was possible, and the time stamps should be used with extreme caution.

*Research Vessel--Shiyan (BXMC)*
Variable	Units Added	Unrealistic Platform Movement	> 4 S.D. From Climatological Mean	Spike	Time Duplicate	Total Number of Flags	Percent of Variable Record Flagged
TIME	674				2	676	100.00
LAT	669	7				676	100.00
LON	669	7				676	100.00
DIR	676					676	100.00
SPD	672		4			676	100.00
P	676					676	100.00
TD	676					676	100.00
RH	676					676	100.00
T	676					676	100.00
TW	676					676	100.00
TS	675			1		676	100.00
Total	7415	14	4	1	2	7436
Percent of Total Data Flagged	99.7176	0.1883	0.0538	0.0134	0.0269	100.00

Most of the data from the bridge observations from the research vessel Shiyan are considered to be in good condition. The data arrived at the DPC without indications of what units were used. Therefore, the units of the variables had to be determined by comparing the data to COARE operational surface maps. The units were added and all of the data was flagged appropriately. Very few flags were added to the remainder of the data.

*Research Vessel--Le Noroit (FITA)*
Variable	Unrealistic Platform Movement	Suspect Data	Spike	Time Duplicate	Total Number of Flags	Percent of Variable Records Flagged
TIME				2	2	0.34
LAT	3				3	0.51
LON	3				3	0.51
DIR		591			591	100.00
SPD		591			591	100.00
P			2		2	0.34
TS			1		1	0.17
Total	6	591	4	2	1193
Percent of Total Data Flagged	0.1128	22.2222	0.0564	0.0376	22.43

Most of the data from the research vessel Le Noroit are considered to be in good condition. The only problem occurs in the winds. A large number of wind speed values were originally flagged as being greater than four standard deviations from the da Silva climatological mean (da Silva, et. al 1994). During inspection, it was discovered that the wind speeds and directions were inconsistent with nearby buoys and ships. As no reason for the inconsistencies was available, the data were flagged as suspect and should not be used. Few flags were added to the remaining data.

*Research Vessel--Hakuho Maru (JDSS)*
Variable	Out of Bounds	> 4 S.D. From Climatological Mean	Total Number of Flags	Percent of Variable Records Flagged
PL_SPD	49		49	2.43
SPD		5	5	0.25
Total	49	5	54
Percent of Total Data Flagged	0.2025	0.0207	0.22

The data from the research vessel Hakuho Maru is considered to be in excellent condition. The only problem was some platform speeds which were between -0.1 m/s and -1.0 m/s and were flagged as being out of realistic range. The cause of this problem is unclear.

*Research Vessel--Franklin (VJJF)--Bridge Observations*
Variable	> 4 S.D. From Climatological Mean	Total Number of Flags	Percent of Variable Records Flagged
SPD	1	1	0.04
Total	1	1
Percent of Total Data Flagged	0.0029	0.00*

*Less than 0.005% of Data Flagged

Most of the data from the research vessel Franklin bridge observations are considered to be in excellent condition. The only problem was found in the wind direction data. After comparing the wind directions to the direction from the research data, a 180 degree difference was found between the two data sets. The directions were corrected by adding 180 degrees to all values (see section 3(c)). Through inspection, it was determined that the time stamps on this data set were in local time and not UTC. After investigation, the correct local time was determined and 10 hours were subtracted from the time stamps to set them to UTC. The time stamps are now correct.

*Research Vessel--Franklin (VJJF)--Research Data*
Variable	Out of Bounds	Unrealistic Platform Movement	> 4 S.D. From Climatological Mean	Platform Crosses Land	Platform Position Uncertain	Spike	Total Number of Flags	Percent of Variable Records Flagged
LAT		4		1	605		610	14.48
LON		4		1	605		610	14.48
PL_SPD	178						178	4.23
SPD			1				1	0.02
TS						2	2	0.05
TS2	3		21				24	0.57
Total	181	8	22	2	1210	2	1425
Percent of Total Data Flagged	0.3581	0.0158	0.0435	0.0040	2.3940	0.0040	2.82

Most of the research data from the vessel Franklin in considered to be in good condition. Position uncertainty during the first cruise was caused by a problem with the GPS receiver so these data should be used with caution. Also, several platform speed values between 0.0 m/s and -1.0 m/s were flagged as being out of realistic range and should be used with caution. Only a few of the remaining data values were flagged.

Research Vessel--Wecoma (WSD7079)

The data from the research vessel Wecoma are considered to be in excellent condition. No flags were added to these data.

*Research Vessel--Vickers (WTEC)*
Variable	Unrealistic Platform Movement	> 4 S.D. From the Climatological Mean	Erroneous Data	Platform Crosses Land	Intercom parison Error	Spike	Total Number of Flags	Percent of Variable Records Flagged
LAT	82			1			83	0.29
LON	82			1			83	0.29
PL_HD						86	86	0.30
PL_SPD			13			27	40	0.14
DIR						38	38	0.13
SPD		43				165	208	0.73
P						249	249	0.88
T		6	622		30	6	664	2.34
TS		148				66	214	0.76
RH			750			48	798	2.82
Total	164	197	1385	2	30	685	2463
Percent of Total Data Flagged	0.0526	0.0632	0.4444	0.0006	0.0096	0.2198	0.79

Most of the data from the research vessel Vickers are considered to be in good condition. Minor problems occurred in the relative humidity and temperature data. Many of these values were very high or low and were flagged as erroneous data.

*Research Vessel--Moana Wave (WUS9293)*
Variable	Unrealistic Platform Movement	> 4 S.D. From Climatological Mean	Erroneous Data	Intercomp arison Error	Spike	Total Number of Flags	Percent of Variable Records Flagged
LAT	24					24	0.30
LON	24					24	0.30
PL_HD					64	64	0.79
DIR					3	3	0.04
SPD		6			19	25	0.31
T					5	5	0.06
T2		1			9	10	0.12
TS			5	9	18	32	0.40
Q					3	3	0.04
Q2					3	3	0.04
Total	48	7	5	9	124	193
Percent of Total Data Flagged	0.0495	0.0072	0.0052	0.0093	0.1279	0.20

The data from the research vessel Moana Wave are considered to be excellent condition. Only minor problems occurred

*Research Vessel--Kexue (BCCU)*
Variable	Out of Bounds	Unrealistic Platform Movement	>4S.D. From Climato log- ical Mean	Erroneous Data	Suspect Data	Platform Crosses Land	Inter comparison Error	Platform Position Uncertain	Spike	Total Number of Flags	Percent of Variable Records Flagged
LAT		591				30		161		782	0.63
LON		591				30		161		782	0.63
PL_HD					425			161	16	602	0.49
PL_SPD	70			390	11			161	134	766	0.62
DIR					123344				51	123395	100.00
SPD	60		68		123180				87	123395	100.00
P									3193	3193	2.59
T			196				31		10	237	0.19
RH					300		906		4	1210	0.98
Total	130	1182	264	390	247260	60	937	644	3495	254362
Percent of Total Data Flagged	0.0096	0.0871	0.0194	0.0287	18.216	0.0044	0.690	0.474	0.2575	18.74

Some of the data from the research vessel Kexue are considered to be in good condition. However, numerous problems occurred in this data set. Problems with the wind data stem from errors in the magnetometer data. A calibration problem with the magnetometer resulted in wind data being suspect. The DPC flagged the wind data accordingly and recommend that the AL Processed wind data for this vessel be used instead (Hartten, 1996). Problems also occurred in the position data, much of the relative humidity data was not consistent with surrounding observations, and many spike flags were added to the atmospheric pressure data. The relative humidity data should be used with caution.

*Research Vessel--Malaita*
Variable	Out of Bounds	Failed T„Tw„Td Test	Erroneous Data	Intercompa rison Error	Total Number of Flags	Percent of Variable Records Flagged
PL_HD	2				2	0.14
T		9	310		319	21.64
TD		9	310	16	335	22.73
RR	18				18	1.22
Total	20	18	620	16	674
Percent of Total Data Flagged	0.1357	0.1221	4.2062	0.1085	4.57

*Research Vessel--Shiyan (BXMC)--ISS*
Variable	Out of Bounds	Unrealistic Platform Movement	>4S.D. From Clima tolog- ical Mean	Erroneous Data	Suspect Data	Platform Moves over Land	Platform Position Uncertain	Spike	Total Number of Flags	Percent of Variable Records Flagged
LAT		4194				99	434		4727	3.03
LON		4194				99	434		4727	3.03
PL_HD					4684		466	163	5313	3.40
PL_SPD	1			4064	566		382	324	5337	3.42
DIR					155953			214	156167	100.00
SPD			88	1892	153932			255	156167	100.00
P				62145				195	62340	39.92
T			13	2218	1115			18	3364	2.15
RH				2154	165			6	2325	1.49
Total	1	8388	101	72473	313599	198	1716	1175	398437
Percent of Total Data Flagged	0.0001	0.4883	0.0059	4.2189	18.4194	0.0115	0.0999	0.0684	23.31

The data from the research vessel Malaita are considered to be in good condition. The only major problem occurred in the temperature and dewpoint temperature data from the early part of the cruise. Many of the values were unrealistically high and were flagged appropriately.

Research Vessel--Shiyan (BXMC)--ISS

Variable Out of Bounds Unrealistic Platform Movement >4S.D. From Clima tolog-
ical
Mean Erroneous
Data Suspect Data Platform Moves over Land Platform Position Uncertain Spike Total Number of Flags Percent of Variable Records Flagged

LAT 4194 99 434 4727 3.03

LON 4194 99 434 4727 3.03

PL_HD 4684 466 163 5313 3.40

PL_SPD 1 4064 566 382 324 5337 3.42

DIR 155953 214 156167 100.00

SPD 88 1892 153932 255 156167 100.00

P 62145 195 62340 39.92

T 13 2218 1115 18 3364 2.15

RH 2154 165 6 2325 1.49

Total 1 8388 101 72473 313599 198 1716 1175 398437

Percent
of Total
Data
Flagged 0.0001 0.4883 0.0059 4.2189 18.4194 0.0115 0.0999 0.0684 23.31

The ISS data from the research vessel Shiyan are considered to be of marginal quality. As with the Kexue, due to errors in the magnetometer data (Hartten, 1996), the DPC recommends that the AL Processed ISS winds for this vessel be used. A calibration error resulted in the wind data being suspect and the DPC flagged the wind data accordingly. Numerous flags were also added to each of the remaining variables.

5. References

da Silva, A. M., C. C. Young, and S. Levitus, 1994: Atlas of Surface Marine Data, Volumes 1: Algorithms and Procedures. NOAA Atlas NESDIS 6, U.S. Dept. of Commerce, National Oceanic and Atmospheric Administration, 83 pp. Available from: National Oceanographic Data Center, User Services Branch, NOAA/NESDIS E/OC21, Washington, DC, 20235.

Hartten, L.M., 1996: Reconciliation of Surface and Profiler Winds at ISS Sites. J. Atmos. Oceanic Technol., Submitted.

Smith, S.R. and D.M. Legler, 1995a: NetCDF Code Manual for Quality Controlled Surface Meteorological Data. Report COAREMET 95-4, Center for Ocean Atmospheric Prediction Studies, Florida State University, Tallahassee, FL 32310.

Smith, S.R. and D.M. Legler, 1995b: ASCII Code Manual for Quality Controlled Surface Meteorological Data. Report COAREMET 95-11, Center for Ocean Atmospheric Prediction Studies, Florida State University, Tallahassee, FL 32310.

Smith, S.R., J.P. Camp, and D.M. Legler, 1996: TOGA COARE: Handbook of Quality Control Procedures and Methods for Surface Meteorological Data. COAPS Report 96-3, Center for Ocean Atmospheric Prediction Studies, Florida State University, Tallahassee, FL 32310.

COARE-MET Flux Algorithm COARE-MET Reports Related Servers COARE Staff

Table 2: Definitions of COARE quality control flags and reference page numbers.
Flag	Definition
A	Original data had unknown units. The units shown were determined using a climatology or some other method.
B	Original data were out of the range bounds outlined.
C	Time data are not sequential or date/time not valid.
D	Data failed T„Tw„Td test. In the free atmosphere, the temperature is always greater than or equal to the wet-bulb temperature, which in turn is always greater than or equal to the dew point temperature.
E	Data failed resultant wind recomputation check. When the data set includes all variables required, a program recomputes the earth relative wind speed and direction and compares the computed values to the reported earth relative wind speed and direction. A failed test occurs when the wind direction difference is > 10š or the wind speed difference is > 5 m/s.
F	Platform velocity unrealistic. Determined with platform position data.
G	Data are > 4 standard deviations from the COADS climatological means (da Silva et al. 1994). Test applied only to pressure, temperature, sea temperature, relative humidity, and wind speed.
H	Discontinuity found in data.
I	Interesting feature found in data. More specific information on the feature is contained in the QC reports. Examples include: hurricanes, sharp sea water temperature gradients, strong convective events, etc.
J	Data are of poor quality by visual inspection, DO NOT USE.
K	Data suspect - USE WITH CAUTION - flag applied when the data look to have obvious errors, but no specific reason for the error can be determined.
L	Oceanographic platform passes over land.
M	Known instrument malfunction.
N	Inconsistent with neighboring station.
P	Position of platform or its movement are uncertain. Data should be used with caution.
Q	Data arrived at the DPC already flagged as questionable.
S	Spike in the data. Usually one or two sequential data values (sometimes up to four values) that are drastically out of the current data trend. Spikes occur for many reasons, including power surges, typographical errors, data logging problems, lightning strikes, etc.
T	Time duplicate.
V	Colocated data differ - use with caution.
Z	Data passed evaluation.

Station Name	Station ID	Number of šF to šC Conversions
Lukunor Atoll	94339	2
Ponape Island	91348	1
Jaluit Atoll/Jabor	91369	1
Majuro/Marshall Island	91376	1
Funafuti Intl Airport	91643	1
Padang/Tabing	96163	2
Biak/Mokmer	97560	1
TOTAL		9

VISIT (Level 2) Quality Control Report for COARE Surface Meteorological Data

(a) (b)

(a) (b)

Buoys--WHOI

Buoys--PMEL

Drifter--Marisonde

Drifter--Niller

Land--CMAN

Land--GTS

Land--Kapingamarangi ISS

Land--Kavieng ISS

Land--Manus ISS

Land--Nauru ISS

Research Vessel--Natsushima (7JDU)

Research Vessel--Xiangyanghong (BNTA)

Research Vessel--Shiyan (BXMC)

Research Vessel--Le Noroit (FITA)

Research Vessel--Hakuho Maru (JDSS)

Research Vessel--Franklin (VJJF)--Bridge Observations

Research Vessel--Franklin (VJJF)--Research Data

Research Vessel--Wecoma (WSD7079)

Research Vessel--Vickers (WTEC)

Research Vessel--Moana Wave (WUS9293)

Research Vessel--Kexue (BCCU)

Research Vessel--Malaita

Research Vessel--Shiyan (BXMC)--ISS