Meteorological buoy measurements in the Iceland Sea 2007 – 2009

The Icelandic Meteorological Office (IMO) conducted meteorological buoy measurements in the central Iceland Sea in the time period 2007–2009, specifically in the Northern Dreki Area on the southern segment of the Jan Mayen Ridge. Due to difficulties in deployment and operations in-situ measurements in this region are sparse. Here the buoy, deployment and measurements are described with the aim of giving a future user of the data set as comprehensive information as possible. The data set has been quality checked, suspect data removed and the data set made publicly available from Pangaea Data Publisher 5 (doi:10.1594/PANGAEA.876206).


Introduction
The Icelandic Meteorological Office (IMO) conducted measurements in the central Iceland Sea in the time period 2007-2009, specifically in the northern Dreki area on the southern segment of the Jan Mayen Ridge.The deployment was a part of a governmental preparation project for the potential exploration for oil and gas on the Icelandic continental shelf and the purpose of the measurements was to obtain information on the local weather.The measurements were conducted with a meteorological buoy making basic in situ meteorological measurements as well as some oceanographic measurements.After the end of the deployment the government phased out the project and the data remained untouched and unused for several years.
Lately the data have become of interest to the scientific community.One reason is the discovery of the North Icelandic Jet (Jonsson and Valdimarsson, 2004), a deepreaching ocean current hypothesized to originate in the Iceland Sea (Våge et al., 2013), although the exact source and related water mass transformation processes are not known.The Iceland Sea is a local heat flux minimum and thus the oceanic deep convection is not driven by the average large-scale atmospheric circulation (Moore, 2012).However, Harden et al. (2015) used the buoy data presented here to examine the surface meteorological condition in the central Iceland Sea and concluded that although on average the heat flux was low, on shorter timescales the Iceland Sea frequently experienced high heat flux events in the wintertime.
The buoy measurements are unique due to the sparsity of in situ observations over the ocean, especially in this region, and Dukhovskoy et al. (2017) used it in an evaluation of ocean surface winds from reanalysis data sets and scatterometer-derived gridded products.
Although measurements from the data set have been used in both Harden et al. (2015) and Dukhovskoy et al. (2017) they have not been made publicly available earlier.However, now the data set has been quality-checked, suspect data removed and made publicly available from the PANGAEA Data Publisher.
The purpose of this article is to give a full description of the buoy, deployment and measurements for any future users of the data set.
In the following section there is a description of the buoy, parameters measured and the deployment.Section 3 contains information on the measurements, the availability of data and, where possible, reasons for periods of missing data.Lastly, final remarks are in Sect. 5.
Published by Copernicus Publications.

The buoy and the deployment
The buoy was a SEAWATCH Wavescan wave buoy from Fugro OCEANOR that measured wave, current and meteorological parameters (Fugro OCEANOR, 2017).The hull was discus-shaped with a keel mounted at the bottom to prevent capsizing of the buoy.The meteorological sensors and the antennae were mounted on a mast (see Fig. 1).The hull had a diameter of 2.8 m and total height (mast to keel) of 6.75 m.It had solar panels and sealed lead acid backup batteries.Due to the low sun radiation condition during winter at the measurement site the buoy was also supplied with lithium batteries.Table 1 lists the instrumentation of the buoy and, for each parameter, range and accuracy according to the manufacturers (Fugro OCEANOR AS, 2007).The buoy was deployed in the northern Dreki area1 on the southern segment of the Jan Mayen Ridge, anchored at 68.47 • N, 9.27 • W from 23 November 2007 to 21 August 2009 (see Fig. 2), drifting inside a circle with a diameter of approximately 2 km.It was serviced once during the deployment, on 7 June 2008.The internal compass and the GPS sensor failed on 17 April 2009 between 08:00 and 09:00 UTC.Thus after that time the exact location of the buoy is not known and there are no measurements of current speed and

The measurements
A list of measurements conducted, units, resolution and short names in the data set can be found in Table 2.The data were recorded every hour.The measurement time period was 23 November 2007 03:00 UTC-21 August 2009 20:00 UTC and Table 3 contains a list of times when some or all of the measurements were missing and the reason when known.In addition other obvious errors were removed, such as spurious longitude = 0 • .Table 4 contains averages for most of the meteorological parameters measured, and maxima and minima where appropriate, while Table 5 contains averages for the main oceanic parameters.The daily values and average, maximum and minimum values are calculated and then the monthly averages of the daily values.In addition to measured values the dew point, T d , the vapour pressure, P v , and the difference between the water temperature and the air temperature are calculated and the mean values included in the table.

Air pressure
Air pressure was measured with a Vaisala PTB220A digital barometer designed to operate over a wide pressure and temperature range (Fugro OCEANOR AS, 2007).A comparison of the pressure data to ECMWF operational analysis confirmed the suspicion that the measurements were off for months after the passing of the deepest low, on 24 October 2008, possibly due to water in the sensor inlet.The error was not just a bias error; in addition, during the first half the pressure was too high, while during the second half the magnitudes of extreme were too small and the extrema were not all accounted for.From 1 February 2009 the measurements were in agreement with ECMWF analysis (not shown).To avoid any confusion the pressure data from 24 October 2008 to 1 February 2009 are removed from the data set.The buoy measurements of mean sea level pressure are shown in Fig. 3, with the exception of the erroneous data.During the deployment, the average pressure was 1009 hPa, varying from a minimum of 941 hPa to a maximum of 1047 hPa.Note that the maximum was measured on 13 February 2008 and a  • bins) at 4 m height.Note that current direction is defined as the direction the current is streaming toward, which is opposite to the convention of wind direction.

Wind speed and wind direction
The wind speed, wind gust and wind direction were measured at 4 m height with a Gill WindSonic wind sensor (Fugro OCEANOR AS, 2007).It has been shown that during rough seas, due to sheltering effects and elevation changes, wind measurement by buoys can be negatively biased (e.g.Large et al., 1995;Zeng and Brown, 1998).Here, no attempt is made to compensate for a potential bias in the data set; that is left to the user.
The maximum measured wind was 19.3 m s −1 and the maximum gust 29.8 m s −1 .The gust factor was in general below 1.5.There was a significant lower wind speed during the summer months than the winter months (see Fig. 5).The  4).
The wind rose in Fig. 6 shows that the most common wind directions were northerly to northeasterly, approximately 30 % of the time, and the least common wind directions were westerly to northwesterly.This is in accordance with wind directional frequency in Iceland (not shown).

Relative humidity
The relative humidity was measured at 3.5 m height with a Vaisala HMP 45A relative humidity sensor based on the capacitive thin film polymer HUMICAP 180 (Fugro OCEANOR AS, 2007).As the measurements were made over the sea the relative humidity was in general high with Earth Syst.Sci.Data, 9, 779-789, 2017 www.earth-syst-sci-data.net/9/779/2017/ Time (s) Spectral peak period the lowest measurement at 50 % (see Fig. 7).In all cases of relative humidity below 55 % the air temperature was below freezing and in most cases winds were light.The mean relative humidity was 87 %.

Water temperature
Water temperature was measured at 1.5 m depth with the Aanderaa DCS 4100R current speed and water temperature sensor.The temperature was measured using a temperaturedependent crystal-oscillator circuit (Fugro OCEANOR AS, 2007).After 17 April 2009 all measurements from the current sensor were invalid.The water temperature measurements were thus made over 16 consecutive months.The lowest temperature was measured in spring and the highest from the end of July until mid-September.The rise in temperature during late spring and early summer was slower than the fall in the autumn (see Fig. 8).The lowest and highest monthly means were in April and August 2008.The air temperature is also shown in the figure.The amplitude of the water temperature variations is much less than that of the air temperature.
The differences between the water temperature and the air temperature, T w −T , were calculated, and on a monthly basis the difference varied between 0.5 and 3.3 • C, with the least difference during summer months but large variation during the winter months, related to different air mass impacting the area.

Current speed and direction
Current speed and direction were measured at 1.5 m depth with the Aanderaa DCS 4100R current speed and water temperature sensor.The sensor used the Doppler shift principle and together with orientation from the internal compass determined current speed and direction (Fugro OCEANOR AS, 2007).As mentioned earlier, after the compass and the GPS sensor broke down on 17 April 2009 all current measurements are missing.As stated in Sect. 2 the buoy was anchored at 68.47  17.2 cm s −1 in February 2008.By convention, current directions are defined opposite to wind directions; northerly current moves toward north while northerly wind is coming from north.The most common current directions were southwesterly (toward southwest) and easterly directions (toward east) least common (see Fig. 10).

Wave parameters
The wave parameters are measured with Fugro OCEANOR Wavesense, a robust integrated wave sensor and data logger.Accelerometers, rate gyros and magnetometers were mounted orthogonally to provide the basic data.These data were then used as input to algorithms, which calculated heave, roll, pitch, surge, sway and compass time series.Significant wave height, wave direction, wave period and a number of other statistical parameters were then found from these time series (Fugro OCEANOR AS, 2007).Figure 11 shows the frequency of wave directions.By convention, wave directions are defined in the same way as wind directions.
The most common directions were northeasterly to east-northeasterly, a slight clockwise rotation from the most common wind directions (see Fig. 6).The least common wave direction was northwesterly. Figure 12 shows the significant wave height (the mean height of the highest one-third of the waves) and the spectral peak period (the wave period with the highest energy) and Fig. 13 the highest wave height and the mean wave period (the mean of all wave periods).Both wave height and wave periods have an annual variation with minimum during summer and maximum during winter, as well as more variability during winter.The monthly mean significant wave height varied from 1.2 to 3.9 m and the monthly mean maximum wave height from 1.7 to 5.9 m.

Data availability
The processed data set and a selection of daily and monthly mean parameters are available from the PANGAEA Data Publisher (https://doi.org/10.1594/PANGAEA.876206).

Conclusions
The Icelandic Meteorological Office deployed a meteorological buoy in the northern Dreki area on the southern segment of the Jan Mayen Ridge for 21 months, from 23 November 2007 to 21 August 2009.This is a region of the North Atlantic with few in situ measurements, and thus the data set is unique.The data set has been quality-checked and is now publicly available.This short paper presents the data set, which parameters were measured and at which height, as well as data gaps.The figures in this paper are meant to give a potential user of the data set a quick view of the data.It is the hope of the author that the measurements can be of use for scientists studying the meteorology and oceanography of the northern North Atlantic as well.

Figure 2 .
Figure 2. The location of the buoy, November 2007-April 2009.The retrieval location on 21 August 2009 is marked with a star.

Figure 4 .
Figure 4. Measurements of air temperature ( • C) at 3.5 m height.A histogram of measurements is inset.

Figure 5 .Figure 6 .
Figure 5. Measurements of wind speed and wind gust (m s −1 ) at 4 m height: time series (a) and histograms of wind speed measurements (b) and of wind gust measurements (c).

Figure 9 .Figure 10 .
Figure 9. Measurements of current speed (cm s −1 ) at 1.5 m depth.A histogram of measurements is inset.

Figure 11 .
Figure 11.A frequency rose, showing the frequency (%) and height of highest wave (m) for different wave directions (22.5 • bins), with wave directions defined in the same way as wind directions.

Figure 12 .
Figure 12.Measurements of significant wave height (m) and the spectral peak period (s).(a) Time series and (b) histograms.

Figure 13 .
Figure 13.Measurements of the highest wave height (m) and the mean wave period (s).(a) Time series and (b) histograms.

Table 1 .
The instrumentation of the buoy: parameters measured, instrument, range and accuracy according to manufacturers.
• Figure 1.The meteorological buoy on board the Icelandic Marine Research Institute's vessel Bjarni Saemundsson on 22 November 2007, the day before deployment.Photo: Sighvatur K. Pálsson.

Table 2 .
Buoy measurements, units, resolution and short names used in the hourly data set.

Table 3 .
List of periods where measurements are missing or clearly erroneous, parameters affected and reasons.
* Pressure measurements were compared to ECMWF operational forecasts.direction.After the failure the buoy broke free.Using the locations of the satellites retrieving information from the buoy it can be seen that this happened in May and the buoy then started drifting northward.It was rescued by the Marine Re-search Institute on 21 August 2009 at 69.40 • N, 9.62 • W, having drifted about 104 km to the north from the mooring location; see location in Fig.2.

Table 4 .
Monthly average values, as well as maximum and minimum values where appropriate, for most of the meteorological parameters.The dew point, T d , and the vapour pressure, P v , are calculated using an equation fromWMO (2012).Insufficient data is represented with -.

Table 5 .
Monthly average values for a few of the oceanographic parameters: water temperature, current speed, mean wave period, height of highest wave, significant wave height and period of spectral peak.The parameter (T w − T ) is the monthly averaged difference between the water temperature and the air temperature.Insufficient data is represented with −.
CountFigure 3. Measurements of mean sea level pressure (hPa).The gap from 28 October 2008 to 1 February 2009 is due to measurement errors.A histogram of measurements is inset.