Water vapor is important to the weather and climate because of their role in the global climate system. The clear interaction between water vapor and atmospheric events can explain some physical mechanisms of how small scale of atmospheric environment could influence the climate change pattern. To enhance this interaction, this work is aimed to observe the characteristics of coreless winter event using precipitable water vapor (PWV) derived from GPS technique, surface meteorology and solar radiation data. The periods of observations are over 2009 for Antarctic and from July 2008 to June 2009 for Arctic. Results show that the occurrence of coreless winter was clearly detected in June and January for Antarctica and Arctic, respectively. During the period of winter, the temperature, relative humidity and PWV variations at both regions demonstrates directly proportional to each other than with surface pressure. At the middle of winter when coreless event took place as indicated by significant unusual warming peak of temperature, relative humidity and PWV in both regions, their pattern was shown similar characters except for surface pressure. During this event, the increasing of 1°C of temperature has increased the PWV of about 0.25 mm and 0.70 mm for Antarctica and Arctic, respectively, verified that the PWV in Arctic was observed twice bigger compared to Antarctic. The increased PWV during winter suggests that the coreless winter characteristic was signifying when advection of warm or cold air masses over the region tend to increase the formation of cyclonic activity that often causes to halt the cooling in surface temperature.
Keywords: GPS PWV; Coreless winter; Antarctic and Arctic; Climate
Introduction
The Antarctic-Arctic regions (bipolar) have growing recognition that polar climate conditions were strongly influence the world climate system. Because of bipolar being in forefront of climate changes issues and a sensitive indicator of global-scale climate change, proper characterization of the polar atmosphere is essential to improve our understanding of the coupling mechanisms between bipolar and global climates, and between the atmospheric, land and oceanic components of the climate system. Accordingly, atmospheric water vapor is particular important because of their capability to regulate the polar energy balances. For example, small changes of atmospheric water vapor have much larger impact on the greenhouse effect and thereby heat the Earth's surface cause a warming. As can be seen, a part of both regions in recent year had shown most rapid rates of the warming impact. In the Arctic, the significant warming commence during the 20th century with magnitude of air temperatures over extensive land areas was expected increased by up to 5°C (IPCC, 2007). The attributions of recent changes are from the natural variability and anthropogenic forcing, which concludes a substantial proportion of the recent variability and manifestation of greenhouse gas induced by human (e.g., Serreze and Francis, 2006). The most recent (1980 to present) warming of the Arctic is strongest (about 1°C/decade) in winter and spring (McBean et al., 2005), while in the Antarctic region, there has been a marked most rapidly warming trend was in the Antarctic Peninsula over the past several decades (Turner et al., 2007). Antarctic Peninsula stations show a consistent regional rate of warming that is more than twice the average for other Antarctic stations. King and Harangozo (1998) suggest that this warming is associated with an increase in the northerly component of the atmospheric circulation over the Peninsula and perhaps changes in the sea-ice extent.
The significant warming trends during winter in Antarctic and Arctic regions have been attracted many researcher to study about the physical mechanism that contributed to the event. The warming temperature during winter in the interior of Antarctica was widely accepted, as the 'coreless' or 'kernlose' winter (see definition by Wexler, 1958; Wendler and Kodama, 1993). This phenomenon refer to the winters without cold core, however the temperature trend increase for a few months and lead to a maximum in early winter after a significant drop in autumn season. Several authors (Carroll, 1982; Stone et al., 1989, and Stone and Kahl, 1991) studied that the increasing of temperature during austral winter was not monotonic and strong sudden warming episodes has been occurred. More studies on temperature behavior during winter in Antarctica, for example, Connolley and Cattle (1994) uses pressure and temperature fields to improve the performance of the UKMO Unified Model and found the coreless feature was present in their model, although cloud covers appears to be a problem to their model accuracy. Van Den Broeke (1998) studied the influence of semi-annual oscillation (SAO) on near-surface temperature in Antarctica for period from 1957 to 1979, and shown sea ice extent modifies the SAO influence during winter, particularly in Vostok station. Styszyńska (2004) examined the relationship between the air temperature at Arctowski station on the South Shetlands, with the sea ice extent and sea surface temperature (SST) in the Bellingshausen Sea, and the event of coreless winter was identified pronounced in July. In addition, Hudson and Brandt (2005) study the relationship between the inversion temperature profile by radiosonde data over the Antarctic Plateau at different scales with winds and downward long-wave radiation. In these studies, however the effect of radiosonde data was limited to demonstrate their temporal resolution because the data only accessible twice a day. Because winters are severe in Antarctica and the period of coreless winter was remarkable different for each region, their origin and causes are important before it strikes. Moreover, the physical processes of coreless during the winter is still poorly understood. Therefore, in this work water vapor monitoring from ground-based Global Positioning System (GPS) technique is proposed to characterize the coreless winter behavior.
The GPS is an accurate and powerful technique to retrieve the precipitable water vapor (PWV) in all weather conditions from single station observations and/or in ground-based network with fine temporal and spatial resolution. This technique was first described by Bevis et al. (1992, 1994), Rocken et al. (1993) and Duan et al. (1996), which GPS satellite radio signals are slowed caused by originates of both the ionosphere and the neutral atmosphere. To determine the PWV from GPS signals delays, the process is accomplished by separating the errors introduced into the calculation by system-related and geometric factors caused the passage of the GPS signal through the atmosphere (Gutman et al., 1994). As the ionosphere delay is frequency dependent, it can be corrected by using dual-frequency GPS receivers, and the remaining delay, neutral delay, which is depend on its constituents in the lower atmosphere. The neutral delay, so-called the total tropospheric delay consists of a 'hydrostatic' delay and a 'wet' delay. The hydrostatic delay containing about 90% of dry gases in the troposphere and the non-dipole component of water vapor refractivity, while the wet delay is associated with the distribution of water vapor overlying from a GPS receiver to the top of atmosphere. On the other hand, the total tropospheric delay is the sum of the hydrostatic delay and wet delay. The PWV can be estimated from the zenith wet delay (ZWD) after the GPS signal mapped to satellite view. Section 2.3 gives briefly explanation of the PWV determination. Currently, the high impact of GPS PWV has been compared with radiosonde or microwave radiometers and found consistent at level of 1~2 mm (Rocken et al., 1997; Businger et al., 1996; Elgered et al., 1998). More presently, it can be used to improve mesoscale NWP model (Kuo et al., 2000) and for climate monitoring (Gradinarsky et al., 2002). The GPS PWV also can be used as a proxy of upper-lower atmospheric coupling studies as proposed by Suparta et al. (2008).
In this paper, the impact of GPS PWV is employed to study the response of coreless winter behavior. In the analysis, the similarities or different of coreless winter in Antarctic and Arctic regions are observed. The location of study for this work is focused at two pairs of polar conjugate stations: Scott Base - Resolute, and Syowa - Reykjavik. Scott Base (SBA) and Syowa (SYOG), and Resolute (RESO) and Reykjavik (REYK) are represents observations for Antarctic and Arctic regions, respectively. For the analysis, the PWV and surface meteorology data over the period of 2009 for Antarctic and from July 2008 to June 2009 for Arctic are processed. The measurements results are then analyzed on a daily and monthly basis to give clear the response of surface parameters on coreless winter event. In further investigation, solar radiation on monthly average is analyzed to get clear seasonal of the occurrence of coreless winters.
Data sets and methodology
Location
In this study the observations of coreless winter were focuses at high latitudes of Southern Hemisphere and Northern Hemisphere, in which the terms of both hemispheres for this work defined as the Antarctic and Arctic, respectively. Two stations in each region, that are SBA and SYOG for Antarctic, and RESO and REYK for Arctic are selected. The location of each station in both regions is located above 60 degrees, and coincidentally identified as geomagnetic conjugate points, whereas the approximate conjugate pairs are SBA - RESO and SYOG -REYK. The advantage of these conjugate stations enabling us to learn many aspects related to the similarities and/or asymmetries of their properties phenomena between the regions. Figure 1 depicts the location of fixed GPS sites at both regions. SBA in the left of figure is located at Pram Point, near the tip of Hut Point Peninsula on Ross Island region within the latitude of 77° 51´S and 166° 46´E longitude. On the other hand, SBA is situated approximately 3 km from McMurdo US base station at Discovery Point, or around 1,353 km from the South Pole. SYOG is also one permanent station managed by the National Institute of Polar Research Japan (NIPR) located on the Ongul Island in Lützow-Holm bay, about 4 km west from the coast of East Antarctica (NIPR, 2009). The station is situated at coordinates 69°00' S and 39°35' E.
On the right panel of Fig. 1 presents the REYK and RESO stations in Arctic region. REYK is located in Southwestern of Iceland, on the southern shore of Faxaflói Bay at geographic: 64.08°N and 21.57°W. The last station, RESO (74.41°N, 94.53°W) is located between the northern end of Resolute Bay and the Northwest Passage of the Qikiqtaaluk region, situated on the Coast of Cornwallis Island in Nunavut, Canada.
Fig. 1. Location of fixed GPS sites in this study for both regions, adapted from http://gdl.cdlr.strath.ac.uk/scotia/vserm/vserm0103.htm
Data sets
The main base of measurement systems for Antarctica is at SBA. The system employed at SBA consists of a GPS receiving system and a ground meteorological system. The GPS receiver was installed in November 2002 under the Malaysian Antarctic Research Programme (MARP) and was maintained by Antarctica New Zealand (ANZ). At this station, GPS data are collected continuously using a Trimble GPS receiver and a Zephyr Geodetic antenna. The GPS receiver was set to track GPS signals at 1s sampling rate and the cut-off elevation angles was set to 13° to eliminate possible multipath effects on GPS data. The surface meteorological data was supported by the National Institute of Water and Atmospheric Research Ltd., New Zealand (NIWA) and ANZ. Both measurement systems housed at Hatherton Geosciences Lab are employed to determine PWV. Further details of measurement systems at SBA can be found in Suparta et al. (2008). The GPS data apart from SBA were obtained from the International GNSS Service (IGS) at SOPAC homepage (http://sopac.ucsd.edu), except GPS data for RESO station, which is from the Canadian Spatial Reference System (CSRS) database. Table 1 gives the instrument setup of GPS receivers and geographical coordinates for four stations in both regions. The GPS data at SOPAC were recorded at 30 second intervals and supplied as RINEX files of 24 hours duration. The all observation files are available in Hanataka format (d-file compression) to reduce the storage size of RINEX files. The surface meteorological data for SYOG, RESO and REYK are obtained from the British Antarctic Survey (BAS), the Environment Canada Weather Office database and the SOPAC websites, respectively. In this work, surface meteorological data for PWV calculation consists of surface pressure (in mbar), temperature (in °C) and relative humidity (in percent). The sampling periodicity for all surface meteorological data for SBA, SYOG, RESO and REYK are 10 min, 3h, 1h and 15 min, respectively.
Table 1
The geographical coordinates and instrument setup of GPS receivers for both regions
Station
(country)
ID
Latitude
(Deg)
Longitude
(Deg)
Height
(m)
Types of GPS receiver and year installed
Cut-off elevation angle (Deg)
Resolute (CAN)
Reykjavik (ICL)
Syowa (JPN)
Scott Base (NZ)
RESO
REYK
SYOG
SBA
74.41°N
64.08°N
69.00°S
77.85°S
94.53°W
21.57°W
39.35°E
166.76°E
34.90
93.01
50.09
15.85
ASHTECH UZ-12 (2006)
TPS E_GGD (2008)
Trimble NETRS (2007)
Trimble TS5700 (2002)
0
0
10
13
To identify a clear seasonal variation of polar climate, solar radiation measured at both regions is employed. The solar radiation data at SBA and Tartu-Toravere (TRV) are obtained from NIWA and the Estonian Meteorological and Hydrological Institute at WRDC (World Radiation Data Center), respectively. The solar radiation consists of direct, diffuse and global, which are measured in unit of W/m2. At SBA, the solar radiation systems are measured with a Kipp and Zonen CM 11 pyranometer to measures direct, diffuse and global radiation. At TRV station, direct solar radiation is measured using the Actinograph, while for diffuse and global solar radiations are measured using the Yanishevsky pyranometer M-150.
Data processing
The PWV total is determined from both GPS and surface meteorological data. As GPS data from SOPAC available in Hatanaka format, to convert a RINEX file to or from Hatanaka format requires special software. This can be obtained from ftp://terras.gsi.go.jp/software/RNXCMP. When RINEX files obtainable in observation and navigation files, the next step is then estimation of total atmospheric delay. Basically there are five steps to derive the PWV from GPS observations. First, the total tropospheric delay is estimated by constraining the positions of widely-spaced GPS receivers and measuring the apparent error in position every 30s. When all system related errors are accounted, the residual error is presumed to come only from the neutral atmosphere. Second, the total signal delays measured by the GPS receiver from all satellites in view are mapped to the zenith direction using a hydrostatic appropriate mapping function, and combined to give the zenith total delay (ZTD). In addition to the precise ZTD estimation accuracy, the residual tropospheric delay was cancelled by implementing a single differencing technique in the pre-processing with baseline length below 10 km. On the other hand, the ZTD in this work is calculated based on the Modified Hopfield model. Third, the zenith hydrostatic delay (ZHD) is calculated using surface pressure measurement and the precise geographic position at the GPS site. In this work, ZHD is used to correct the errors caused by atmospheric delays on the GPS signals. Fourth, the Zenith Wet Delay (ZWD) is obtained by subtracting the ZHD from ZTD. Finally, PWV is derived from ZWD signals and a conversion factor that proportional to the weighted mean surface temperature. The mean air temperature is currently estimated from a surface temperature measurement at the site. In this work, the Tropospheric Water Vapor Program (TroWav) written in Matlab developed by the first author was used to process and analyze all the above parameters. Further details of GPS derived PWV above can be found in Suparta et al. (2008). For this work, the actual PWV data (in kg/m2 or millimeter) at SBA has been calculated at a 10-min interval. The ZTD product at this station had an accuracy of about 1.0 ~ 1.20 cm level, which corresponds to 1~2 mm in PWV bias.
To observe annual oscillation of winter season for both regions, the data collected from January to December 2009 and from July 2008 to June 2009 for two stations in Antarctic and Arctic, respectively, were processed. Overall, the GPS PWV results at SBA, RESO, REYK and SYOG for this analysis are calculated at 10 min, 1h, 2h and 3h intervals, respectively. In further investigation, the monthly average of all parameters with solar radiation was analyzed in order to get clear pattern of coreless winter behavior. The analyses are to observe the temporal variations of surface meteorological measurements and PWV to correlate their influence on the coreless winter.
3. Results and discussions
To observe the characteristics of coreless winter between the regions, the atmospheric variables: surface meteorological measurements (pressure, temperature and relative humidity), PWV and solar radiation components (direct, diffuse and global) during the winter are presented. The dynamic responses of coreless winter on the above variables are discussed through in the physical processes at the origin of these coreless.
3. 1 Daily surface meteorological and GPS PWV variations between the regions
Figure 2 shows the daily variations of surface parameters for Antarctic and Arctic, respectively. The top panel of Fig. 2 presents the surface pressure for SBA and SYOG vary from 939 mbar to 1029 mbar. At each station, their mean values are recorded around 991 mbar and 985 mbar, respectively. For RESO and REYK, they have similar mean values that are about 1000 mbar. The lowest and highest pressures for Arctic are recorded in January and April with values 950 mbar and 1035 mbar, respectively. As shown in the figure, the surface pressure variations at both regions exhibits highly irregular and their amounts are depending on the weight of atmospheric mass on top of particular measurement point. The air pressure at sea level normally varies between 970 mbar and 1040 mbar (Sing and Aung, 2005) with standard atmospheric pressure at sea level is taken as 1013.25 mbar. The middle panel of Fig. 2 presents the daily variations of surface temperature. The average values for Antarctic ranging from -50°C to 4°C with mean values for SBA and SYOG are about -19°C and -10°C, respectively. The range temperature for REYK was between -9.9°C and 24.4°C with a mean value of about 5°C. At RESO, the range values vary from -40°C to 18°C (-14.6°C, on average) with an extreme minimum value recorded was -40.2°C in March. From temperature recorded, temperature for Arctic demonstrates a large amount of heat compared to Antarctic temperature with difference mean value was 10.31°C. The bottom panel of Fig. 2 presents the variations of relative humidity for each station with mean values for SBA and SYOG are about 63% and 73%, respectively, and for REYK and RESO are 77% and 71%, respectively. The relative humidity at REYK shown small fluctuations compared to RESO, whereas the humidity at RESO demonstrated increase from July to November and started decrease in December. The relative humidity in Antarctic was more fluctuating compared to Arctic variation because of the driest atmosphere in Antarctica. Overall, the daily variation of surface meteorological parameters at both regions showed a similar seasonal variation, which highest during summer and lowest for winter periods.
Fig. 2. Daily average of surface meteorological variations for two stations in both regions, respectively. The month label in the figure represents middle of the month in UT.
Figure 3 presents the GPS PWV variations for Antarctic and Arctic, respectively. The daily pattern of PWV in both regions showed U-distribution, which minimum in winter and maximum in summer. This suggests the developing of PWV variability was dependence to the Sun. As shown on the left of Fig. 3, the average values of PWV measurements observed at Antarctic ranging from 0.35 mm to 11.54 mm, with mean values are 3.27 mm and 5.26 mm for SBA and SYOG, respectively. While for Arctic on the right of Fig. 3, the average PWV values vary from 0.75 mm to 31.75 mm, with mean values are 5.42 mm and 12.35 mm for RESO and REYK, respectively. The PWV variation at REYK shown a highly variable compared to the PWV at RESO. This high fluctuations is possibly because its location that tendency have stormy weather influenced by the battle of the Irminger Current and East Greenland Current (see Nowotarski et al., 2006). Thus the PWV value in Arctic was observed approximately twice bigger compared to the PWV at Antarctica. In addition to the weather condition, the wind speeds at SBA and SYOG were windiest, with annual averages of about 5.2 m/s and 6.5 m/s, respectively. The wind direction for SBA, the most frequent wind direction coming from easterly to Northeast between 45° and 90° (see Suparta et al., 2009). For SYOG, the wind dominantly blows northeasterly (southwestward) with directional constancy of wind about 0.78 (Sato and Hirasawa, 2007). While annual average wind speeds for RESO and REYK were about 5.9 m/s and 4 m/s, respectively. Both wind directions at both stations from Southeast to Southwest directional (see Einarsson, 1984) and flow from North to Northwest (NW) directional (http://www.theweathernetwork.com), respectively.
Fig. 3. Daily average of GPS PWV variations at both regions
Coreless winter analysis
3.2.1 Solar radiation pattern between the regions
As introduced in Section 2.2, solar radiation for this work is employed to identify a clear seasonal variation of polar climate in both regions. Because of lack availability of recent solar radiation data at RESO, solar data at TRV station (58°15'N, 26°28'E) for period from July 2008 to June 2009 was chosen instead of RESO data. Although solar data from TRV used in this work, solar data at RESO (http://wrdc.mgo.rssi.ru) for the period of 1998/1999 was compared to TRV, and shown that the annual averages for diffused and global components for RESO was about 585 W/m2. The lower solar radiation recorded at RESO for 1998/1999 with approximately zero values was received from November to February. While average solar radiation at TRV at the same period was recorded lower between October and April for diffuse and direct components about 467 W/m2. In addition, referring to the Earth receives a total amount of radiation in one cross area (one-fourth) about 342 W/m² of solar constant, the occurrence of dark (winter period) for Arctic can be assumed occurs between October and April. The low solar radiation patterns for diffuse, direct and global components recorded at SBA are employed to represent winter period in Antarctic.
Figure 4 shows the variations of solar radiation components at SBA and TRV over the period of 2009 and from July 2008 to June 2009, respectively. On the left of Fig. 4 shown that SBA received low solar radiation around six consecutive months from March (polar sunset) to October (polar morning) with nearly about zero values. Thus the solar radiation pattern in Antarctic represented by SBA, which the period of winter was defined between March and October. While at TRV, solar radiation was recorded low at around five months which considered below 500 W/m2 occurred from October to April. As shown in Fig. 4, solar radiation penetrates higher in Arctic compared to Antarctic, with average of global components at TRV larger about 33 times than with SBA. Similar to SBA, solar radiation at TRV show increasing in spring and maximum in summer, whereas in autumn the variations were declining. The average differences of direct, diffuse and global components during winter between both regions are about 300, 216 and 309 W/m2, respectively. On the other hand, the variations of solar radiation in both regions were characteristically U-distribution. From the low amount of solar radiation in Fig. 4, the assumption for the period of winter in both regions used to analyze the characteristics of coreless winter can be properly expected.
Fig. 4. Monthly average of solar radiation variations at both regions
Characteristics of coreless winter
To monitor the coreless winter behavior in both regions, the surface parameters (pressure, temperature and relative humidity) and PWV variations were analyzed in a monthly average, as presented in Fig. 5. As shown in the figure, the significant unusual warm temperature peaks noticeably occurred during winter was identified as a coreless winter event. For example, the coreless winter event pronounced in Antarctica occurred when the surface temperature has a significant peak for one or two months during winter season (Wexler, 1958; van Loon, 1967; King and Turner, 1997; Styszyńska, 2004). To observe clearly the characteristics of coreless winter, the analyses are focused during the period of winter as indicated by two episodes as F1 and F2. The first episodes (F1) were from March to June of 2009 and October 2008 to January 2009 for Antarctic and Arctic, respectively. The second episode (F2) for Antarctic and Arctic were in June to October 2009, and from January to April of 2009, respectively.
The left panel of Fig. 5 showed the coreless winter behavior in Antarctic started from March to October. June is considered as a coreless winter because of unusual slight warming during the period of winter. In the first episode (F1), the surface pressure variations at both stations have shown gradually increased with peaked in May. The temperature and PWV variations were in declining phase, through relative humidity at SBA was peaked in April. However, there was significant warm peak of temperature at SBA in April through the beginning of winter time, together with the rising peak in PWV, which the increase values of about 0.83°C and 0.32 mm, respectively. Thus the temperature, relative humidity and PWV were drop together in May before it peaked in June. From May to June, when surface pressure gradually decreased to 992 mbar on average, the averages of temperature, relative humidity and PWV shown gradually increased, with gradient values of about 5°C, 4% and 1 mm, respectively. However at SBA, relative humidity was shown decreased in June. Further, one or more can summarized during the F1 episode, first, the averages of temperature and PWV variations exhibit semi-seasonal oscillation, or opposed to the surface pressure variation, while relative humidity reveal a sinusoidal oscillation. Second, the Antarctic is starts cooling, although temperature and PWV shown increased about 4.6°C and 1.02 mm, correspondingly. On the other hand at this episode the cyclonic activity is weak, since the meridian variation in temperature in the middle troposphere is reduced.
In the second episode (F2), both stations had shown decreasing in surface temperature, relative humidity and PWV variations from June to August, which correspond to the continuous of winter season. However, the pressure variation at both stations showed a difference pattern. The pressure in SBA showing increasing from June to July and then starts to decrease, while SYOG pressure was drop from June to July but the variation started increasing in August. After that, the temperature and PWV variations begin increased during late winter and rapidly rises from September to October, while surface pressure variation tend to decrease slowly. On the other hand, the relative humidity variation at SBA showed a significant drop in September and then increased back in October. The increasing of temperature was continuously until December/January in summer, when the Sun is rotating the horizon continuously at about 20° elevation (Warren and Town, 2009). During the F2 episode, it wholly noted that the temperature and PWV variations shown closely related to each other than to surface pressure, while relative humidity was dropped in September. The Antarctic weather in F2 episode is shown colder than F1 episode due to sea ice reaches its maximum extent at the end of the winter. The September-October is considered as an earlier Austral spring in which warming is faster than the autumnal cooling because the formation of the sea ice that takes longer time. On the whole, the coreless winter for Antarctic on the left of Fig. 5 was clearly occurred in June, with the difference in monthly average during the period of winter respective to the peak value of coreless winter for all observation parameters are increased of about 3 mbar, 3.4°C, 1.6% and 0.65 mm, respectively. Figure 6a summarizes the responses of surface parameters and PWV observations during each episode of coreless winter event.
For Arctic, the period of winter is identified between October and April as shown on the right panel of Fig. 5. January is considered as a coreless winter, though the average of surface pressure was significant drop in their month due to depressions of circumpolar vortex. In the F1 episode, the average of surface pressure was increasing in November, and the averages of temperature, relative humidity and PWV was decreasing simultaneously. At surface pressure peak in November, the temperature, relative humidity and PWV at REYK was shown increased than at RESO. Then these three parameters drop together in December, and at the same time surface pressure decrease gradually to 1000 mbar and 996 mbar for RESO and REYK, respectively. From December to January, the three parameters were shown a slightly increased, which is indicated as a warming at the middle of winter. As can be seen at RESO, the temperature and PWV decrease significantly compared to REYK. REYK showed a slightly difference in PWV, which the variation was gradually increased of 0.22 mm. This due to the interaction of North Atlantic Current with other Ocean Current that may bring strong storms and it would influence the rate of precipitation. During coreless winter in January, both stations had increased in temperature and relative humidity, when the pressure was decreased. Thus the warm temperature of about 2.2°C during winter tends to force the PWV variation to increase to 1.83 mm. In summary, the averages of temperature, relative humidity and PWV demonstrates decrease faster, drop in December before gradually increased in January. Surface pressure in this episode was peaked in November and their trend shown opposite to F1 episode in Antarctic. The relative humidity variation at both regions shown crossed each other in the F1 episode during early winter which Antarctic in April and Arctic in November was due to the transition summer to winter season.
In the F2 episode, variation of pressure at both stations had a similar pattern which increased in February. At the same time the temperature, relative humidity and PWV profiles at both stations had shown decreasing. Then, the surface pressure variation for both stations had shown difference pattern. The pressure at RESO demonstrated increasing from February to April, while pressure at REYK shown significantly drop in March and afterward increased back in April. The surface temperature, relative humidity and PWV variations at REYK demonstrated increasing from February to April, whereas at RESO these three variations continuous decreasing until March but increased back in April. From March to April, all the parameters at both stations demonstrated increasing in their values, which correspond to the seasonal transition to summer. For F2 episode, we can summarizes that the surface pressure was opposed variation to the temperature, relative humidity and PWV profiles. The change of temperature and relative humidity in Arctic demonstrated more decreased in F2 than F1 which same as in Antarctic region. The coreless winter for Arctic on the right of Fig. 5 was shown clearly occurred in January at REYK, with the difference in monthly average during the period of winter respective to the peak value of coreless winter for all observation parameters were increased of about 9 mbar, 0.9°C, 2.5% and 0.40 mm, respectively. However, RESO shown less significant warming of temperature and PWV peaks during this winter period due to the sea ice covered through the year and thus the continuous cooling was relatively high than warmer rate.
Fig. 5. Monthly averages of surface meteorology and PWV variations for both regions. The interval time between vertical dashlines show the case for the period of winter, which both divided into two episodes of coreless event, and the vertical dashlines in June and January show a significant slight warming in the middle of the winter (called "coreless winter"). Note that error bars in each graph show the confidence intervals of data or the deviation along a curve, which is obtained from average between the two values at both stations.
The characteristics of coreless winter induce to the surface meteorology and PWV parameters are shown in Fig. 6. As shown in Fig. 6a, in the F1 episode the surface pressure are increased for all stations, however, the temperature and relative humidity had shown similar pattern to each other, which increased at all stations except for SYOG. On the other hand, PWV demonstrated increased for SBA and REYK, and decreased for SYOG and RESO. In the F2 episode, the surface pressure at all stations shown similar pattern as pressure in F1 episode. However, the temperature, relative humidity and PWV profiles shown decreased between the average and peak values for each station. Based on the characteristic of all parameters during both episodes, we can conclude that the coreless winter was significantly induced during F1 episode. Figure 6b summarizes the significant response of coreless winter for Antarctic and Arctic during the period of winter with respect to coreless winter phenomenon. As shown in the figure, the surface pressure at both regions was shown opposite variation to each other which coreless winter peaks in Antarctic shown increasing and Arctic demonstrated decreasing. Average values of all parameters for pressure, temperature, relative humidity and PWV during winter are about 989 mbar, -20°C, 67% and 3.0 mm for Antarctic which are all lower by about 12 mbar, 9°C, 4% and 3.1 mm for Arctic, respectively. Comparing to their peak values, the averages of temperature and PWV for Antarctic are observed increased of about 4°C and 0.36 mm for SBA and 2.8°C and 0.93 mm for SYOG, respectively. While for Arctic, the significant warming had increasing the temperature of 0.92°C and PWV of 0.41 mm for REYK, in contrast there is less response at RESO. Relative humidity was increased only at SYOG and REYK for about 4.9% and 2.5%, respectively. The temperature, relative humidity and PWV profiles had shown similar characteristics, which the coreless winter peak values are higher than their average value during the whole winter, except at RESO the coreless winter peak was lower than the average value. The less response of all parameters at RESO in particular surface temperature and PWV perhaps continuous cooling that may slow the warming. On the occurrence of coreless winter, much climate at RESO is moderated by the ocean water, which is relatively warm water keeps the North Pole from being the coldest place. These indicative factors suggest that why climatic response at RESO during coreless winter event was different compared to other stations.
Fig. 6. Response of surface parameters and PWV during (a) each episode and (b) the period of winter at both regions respect to the coreless winter event. Average and peak in the legend stands for average and peak values during each episode and the average period of winter and peak coreless, respectively.
4. Conclusion
This paper success addressing the characteristics of coreless winter and their impacts on the GPS PWV variability over Antarctic and Arctic regions. The observations carried out on short-term during winter for the period of 2008/2009 showed that June and January months were identified as the occurring of coreless winter event for Antarctic and Arctic, respectively. During the period of winter in both regions the temperature, relative humidity and PWV demonstrates directly proportional to each other than with surface pressure. Temperature and PWV are the two parameters had bigger impacts in the slightly warming of the atmosphere at the middle of winter than with relative humidity. At these coreless winter events, roughly, an increasing in 1°C of temperature will increasing of PWV content of about 0.25 mm and 0.7 mm for Antarctic and Arctic, respectively. In other words, the difference PWV values between Antarctic and Arctic has been quantified at approximately twice bigger compared to Antarctic, as the temperature in Arctic was warmer than with Antarctic. There are similarities characteristics of atmospheric variables on the coreless winter event characterized during winter in both regions. Coreless winter was signified when the significant warming in temperature during winter season related to the warm air advections to the difference temperature between the sea ice cover and sea surface temperature. The circulation will occur when the warmer water surface flows over the cool air surface; hence the heat exchange will increased the air temperature and water vapor. In addition, the increasing of temperature and PWV may be affected by the increasing of mixing warmer or cold air mass advection from the strong wind speeds that will intensify cyclone activity and the travelling cyclones will bring precipitations as noted by Mahesh et al. (2003) and Einarrson (1984).
The analysis can be concluded that the increasing of PWV during winter in both regions signature the GPS signals had slightly delayed that may problem in positioning application. Although a small increase of PWV during winter because of no or very low solar radiation received in that region, its phase would be very significant in order to enhance the capability of ground-based GPS as a remote sensing tool to monitor atmospheric water vapor. The PWV variability in both regions was closely follows the temperature patterns which are clearly identified a seasonal signal; highest in summer and lowest in winter, and unusual warming at the middle of winter is known as the coreless winter. The characterization of the response of surface meteorology and PWV parameters during winter showed that PWV data has more capable and consistent to detect the period of coreless winter, instead of conventional assessment method from the inversion of surface air temperature. However, the prospective PWV data for climate studies is not complement to traditional measurements. Although the time occurrence of coreless winter may different in specific region due to influence of the air circulation factors, the high temporal and continuous GPS PWV data presented here has a great potential to monitor the characteristics of the atmosphere in particular in sensible areas as the poles. To exploit comprehensive physical mechanism of water vapor distribution during coreless winter, new factors and with different event that may influence its circulation employed for mitigating climate change prediction are considered for further studies.
