Killingtveit.indd 161Killingtveit et al. 2003: Polar Research 22(2), 161–174 Hydrological data from Arctic regions are sparse, and Svalbard is no exception. There were, for example, no regular monitoring stations for river runoff before 1990 and even today there are fewer than fi ve stations with continuous record- ing. Water resources can easily become a limiting factor for development of industry and tourism, and knowledge about hydrological processes con- stitute an important link between global change scenarios and consequences for the Arctic eco- systems. Both for scientifi c studies and water management, it is important to know the amount of water in storage in the catchments (as glaciers, snow, groundwater and lakes) and the fl ux of water into or out of the catchments as precipita- tion, evaporation and runoff. Water balance or water budget calculations is a method used by hydrologists to assess the water resources within an area, and to verify that the different measurements of hydrological terms are consistent and gives results that are correct. The water balance concept is based on the fact that during some time (a day, a month, a year, etc.) the total input of water to an area, for example Water balance investigations in Svalbard Ånund Killingtveit, Lars-Evan Pettersson & Knut Sand This paper reviews and summarizes all known previous water balance studies in Svalbard. An updated water balance computation was then done for the three water catchments with the best data: Bayelva, De Geer- dalen and Isdammen/Endalen for 10 hydrological years 1990–2001. The computations were based on the best available data and correction meth- ods. Special emphasis was put on correction of precipitation data, both for catch errors and gradients in precipitation. Areal precipitation in the three catchments is more than two times the measured precipitation at the clos- est meteorological station: 548 mm/year in De Geerdalen, 486 mm/year in Endalen/Isdammen and 890 mm/year in Bayelva. Compared to this, average measured precipitation is only 199 mm/year at Svalbard Air- port, close to Endalen/Isdammen and De Geerdalen, and 426 mm/year in Ny-Ålesund, close to Bayelva. Evaporation is not well understood in Svalbard; the best estimates indicate an average annual evaporation of ca. 80 mm/year from glacier-free areas, and no net evaporation from glaciers. Glacial mass balance has in general been negative in Svalbard during the last 40 years, leading to a signifi cant contribution to the water balance, on the order of 450 mm/year on average. Annual runoff ranges from 545 mm in Endalen/Isdammen, 539 mm/year in De Geerdalen up to 1050 mm/year in Bayelva. Runoff computed from water balance com- pares well with observed runoff, and average error in water balance is less than ± 30 mm/year in all three catchments. Å. Killingtveit, Dept. of Hydraulic and Environmental Engineering, Norwegian University of Science and Technology, NO-7491 Trondheim, Norway, aanund.killingtveit@bygg.ntnu.no; L.-E. Pettersson, Norwegian Water Resources and Energy Directorate, Box 5091 Majorstua, NO-0301 Oslo, Norway; K. Sand, Statkraft Grøner AS, Box 331, NO-7403 Trondheim, Norway. 162 Water balance investigations in Svalbard a river catchment, will be balanced by the total amount of water leaving the area and changes in storage within the area. No water can disappear since it has to be accounted for in the budget. The water balance can, in principle, be calculated for any area, large or small, but in practice the water balance is usually calculated for a river drainage basin, also called a catchment. The usual formulation of the water balance equation for a river catchment is: PA – QS – QG – EA ± ∆M = ε , (1) where PA is precipitation input (mm), QS is sur- face (river) runoff from the catchment (mm), QG is groundwater runoff from the catchment (mm), EA is evaporation from the catchment (mm), ∆M is changes in water storage within the catchment (mm) and ε is an error term (mm). The error term ε should approach 0 if all the fi ve terms on the left side of the equation have been determined cor- rectly. The magnitude of the error term indicates the accuracy of the different terms of the water balance. The type of errors and their magnitude will be different in different hydrological regimes, depending on climatic, geological and topograph- ical conditions, and the quality of measurements. In Arctic catchments many specifi c problems are encountered in the measurements. Methodology—special problems in Arctic catchments The problem of water balance computations in Arctic catchments is illustrated in Fig. 1. Here, the measured runoff in Bayelva (converted to mm) is compared to the measured precipitation at the nearest precipitation station at Ny-Ålesund (see Fig. 2). The comparison is done for 12 hydrologi- cal years (a hydrological year is from 1 October to 30 September) with simultaneous observations of runoff and precipitation. This is a striking exam- ple of what is sometimes called the “hydrologi- cal paradox”. The paradox is that runoff seems to be much larger than precipitation. In reality, it should, of course, be the other way around, since all river runoff is ultimately generated from pre- cipitation, and there will also be evaporation and possibly groundwater runoff removing some of the precipitation from the catchment. The para- dox can be attributed to a combination of meas- urement errors, non-representative location of precipitation stations and net glacial melt during the period of study. By identifying and correct- ing these problems it is possible to establish the true value of each component in the water bal- ance, and fi nally check the water balance in order to verify the computations. The computation of each individual term in the water balance, and some particular problems specifi c for Arctic cli- mate are described below. Precipitation The term PA in the water balance is the average or areal precipitation in the catchment. This term is based on observed (measured) precipitation, but it will usually be higher than measured precipita- tion since precipitation measurements are affect- ed by a number of error sources. In addition, precipitation gauges may not be placed in rep- resentative locations in the catchment, compared Fig. 1. Observed precipitation and runoff in Bayelva. 163Killingtveit et al. 2003: Polar Research 22(2), 161–174 to the areal precipitation variation. One particu- lar problem of great importance is that precipita- tion usually increases with elevation, while most precipitation gauges are located in the lowlands. This is a highly signifi cant problem in Svalbard where all precipitation stations are located close to sea level, while catchments may reach up to more than 1000 m a.s.l. The computation of areal precipitation in gen- eral is based on measurements in one or prefer- ably several gauges located within or near the catchment. The computation usually consists of the following two steps. (1) The observed pre- cipitation (PO1, PO2, … PON) is corrected for gauge catch errors to obtain true point precipitation values at the site of the precipitation gauges (PT1, PT2, … PTN). (2) Areal precipitation (PA) can then be computed as an areal average for the catch- ment. This is based on the corrected point precip- itation data: PA = f (PT1, PT2, … PTN). The method Fig. 2. Location of the three selected catchments in Spitsbergen, Svalbard: Bayelva, Endalen/Isdammen and De Geerdalen. 164 Water balance investigations in Svalbard for areal averaging must consider both regional and elevation gradients in precipitation distribu- tion within the catchment. It is well known that precipitation measure- ments—in particular the measurement of snow precipitation—are subject to large errors (see, e.g. Strutzer 1965, Bogdanova 1968, Larsson & Peck 1974 and Sevruk 1982). The measure- ment error nearly always leads to an observed precipitation that is less than the true point pre- cipitation. Correction methods for gauges in the Nordic countries have recently been devel- oped and reported by Førland et al. (1996). Error sources and correction methods were studied in the Climate Studies in the Arctic project, where one of the sub-projects was to investigate the difference between measured and true precipi- tation at Svalbard. Results from this study and other relevant studies are reported by Hanssen- Bauer et al. (1996). They conclude that “The sea- sonal ratio between true and measured precipita- tion varies between 1.26 for the summer and 1.70 for the winter. If it is supposed that the seasonal ratios which were found are typical for a normal year in Ny-Ålesund, the true normal (1961–1990) would be 550 mm, i.e. 50 % higher than the offi - cial uncorrected value” (p. 37). Water balance calculations by Sand & Bruland (1999) used the correction factors found by Hans- sen-Bauer et al. (1996): 1.26 in June–August, 1.45 in September–November, 1.70 in December–Feb- ruary, and 1.57 in March–May. These correction factors were applied for the precipitation gauges both in Ny-Ålesund and at Svalbard Airport. Elevation gradients in precipitation In mountainous catchments one usually fi nds that precipitation increases with increasing elevation, at least up to some elevation levels. In the Nordic countries it is quite common to observe precip- itation gradients of 5 - 10 % increase per 100 m increase in elevation. In Svalbard all precipitation stations are locat- ed near sea level, and therefore the measured pre- cipitation here can be much lower than the aver- age (areal) precipitation in the catchments which may extend up to elevations of 500 - 1000 m a.s.l. or more (Fig. 3). This problem was recognized when the fi rst attempts on water balance calcu- lations were planned from 1990 onwards. The fi rst observational network for the study of pre- cipitation gradients was set up in De Geerdalen and close to Longyearbyen. The average gradi- ents found for precipitation during summer were 5 - 10 %/100 m. The true gradient may have been higher though, since the gauges located at higher elevations were more exposed to wind than those in lower and more sheltered areas (Killingtveit et al. 1994). At Ny-Ålesund the orographic precipitation distribution was studied during the 1994 and 1995 summer seasons (Førland et al. 1997a). A gradient of 20 %/100 m was found for summer precipitation, at least for elevations up to 300 m. This led to the conclusion that “The orographic precipitation enhancement, and catch defi ciency of conventional precipitation gauges may fully explain the apparent discrepancy between pre- cipitation measured at Ny-Ålesund and runoff/ mass balance estimates for the Bayelva catch- ment”. This result corresponds well with Repp (1979), who found a gradient of 95 mm/100 m (ca. 20 %) and Hagen & Lefauconnier (1993, 1996) who used 25 %/100 m but who also point- ed out that a linear gradient of 25 %/100 m might give too high values in the uppermost areas. Mer- cier (2001) reported average gradients of 100 mm/ 100 m in a small catchment near Zeppelinfjellet, which translates to ca. 12 %/100 m. Sand & Bruland (1999) used a gradient of 25 %/ 100 m in the Bayelva catchment, with reference to Hagen & Lefauconnier (1993). In the Lon don elva catchment they used a gradient of 31 %/100 m, a value found from analyses of snow survey data from the years 1996–98 in the same catchment. Fig. 3. Hypsographic curve for the three selected catchments. 165Killingtveit et al. 2003: Polar Research 22(2), 161–174 Finally, they used a gradient of 14 %/100 m in the De Geerdalen catchment with reference to Kill- ingtveit et al. (1994). Precipitation gradients during winter cannot easily be studied by direct precipitation meas- urements due to the large problems of meas- uring snow precipitation in higher and more wind-exposed locations. Since snow in Svalbard usually accumulates without any signifi cant melt events during the winter, it is possible to use snow on the ground as a measure for winter precipita- tion. This requires a careful selection of measure- ment sites to avoid problems due to wind redistri- bution of snow on the ground. Tveit & Killingtveit (1994) reported gradients of 85, 57, 20 and 85 mm /100 m for the four years 1991–94, based on anal- ysis of snow surveys. This translates into relative gradients from 4 to 28 %/100 m, with an average around 14 %/100 m. Winther et al. (1998) studied regional distribution of snow in Svalbard as well as the elevation gradients. Their study was sup- plemented by Sand et al. (2003), who updated the results based on three years of snow surveys. The average gradient was 97 mm/100 m for the eleva- tion range 100 - 1000 m a.s.l., which translates to 16 %/100 m. Runoff Runoff is usually the most reliably measured term in the water balance. Since runoff is an inte- grated response from the whole catchment, the problem of representativity of stations does not exist for this term. Still, there are other impor- tant problems and error sources, in particular in an Arctic catchment. Practically all continuous runoff measurements are done by calibrating a rating curve which shows the river fl ow (runoff) as a function of river water level (stage). The stage can fairly easily be recorded and the runoff computed as long as the rating curve is valid. Most errors in runoff records are related to prob- lems in establishing calibration and the stability of the rating curve, while the stage measurements in itself is relatively simple to perform. Two of the most important problems sources in Arctic rivers are unstable bed profi les at gauging stations and ice and snow blocking the river profi le. In Svalbard, river runoff mostly occurs during a few months from June up to September. In the autumn all rivers freeze up completely, except short reaches of rivers fed by springs or in the front of some glaciers (Petterson 1994). During the freeze-up process, large amounts of bottom ice may be formed, blocking the river profi le. Later, during the winter, snow is swept down into the river channel and thick snow drifts may form, especially in deep gorges and ravines. When the snowmelt starts in the spring or early summer, the runoff often increases quite rapidly, fl owing into a river channel which is still blocked by ice and snow. In perennial rivers similar problems may occur when surface ice breaks up and creates ice- runs and ice-jams in the river. In all these cases, the water level in the river, which is the property measured at gauging stations, may increase much more than the fl ow. The problem will gradually disappear as snow and ice is eroded and melted, but its magnitude and duration may be very dif- fi cult to determine, especially at remote stations with automatic measurements. Evaporation Evaporation from a catchment occurs from water surfaces, soil surface and vegetation. The total evaporation from the catchment is called areal or actual evaporation. Areal evaporation is deter- mined partly by the climate (potential evapora- tion) and partly by the wetness conditions in the catchment. In a dry desert, the potential evapo- ration may be very high, but the actual evapora- tion will be zero if there is no water. In Arctic catchments the potential evaporation during the summer may be very signifi cant, due to high net radiation during days with 24 hours of sunshine. The amount of actual evaporation may still be small due to little vegetation, little rainfall and soils that easily dry up after snowmelt. Poten- tial evaporation can be determined by measure- ments or by computations based on climatic data. Actual evaporation is usually computed as a fi xed percentage of potential evaporation or as a func- tion of potential evaporation and soil moisture conditions. It may also be computed as a residual term in a water balance computation. Storage change The most signifi cant storage components gen- erally considered in the water balance calcula- tions are: ∆M = ∆MS + ∆MG + ∆ML + ∆MR , (2) where ∆MS is change in snow storage, ∆MG is change in glacier storage, ∆ML is change in 166 Water balance investigations in Svalbard lakes and river storage, ∆MR is a residual term, including change in soil- and groundwater stor- age. Snow storage, lake and river storage and the residual term will normally show limited varia- tions on a seasonal time scale. Glacier storage, however, normally show variations over a much longer time scales and the ∆MG term may become very signifi cant over a period of a few years. If the water balance calculation is done for a long time period (several years) most of these terms can usually be neglected, with the exception of ∆MG. Also, if the water balance is done for hydro- logical years (e.g. 1/9–31/8 or 1/10–30/9) the stor- age terms (except ∆MG) can usually be neglected. For shorter time steps (month, week, day) all the storage terms must be included in the calculation, which complicates the calculations considerably. Previous work Some of the elements in the water balance have been measured for a long time in Svalbard. Pre- cipitation measurements have been made by the Norwegian Meteorological Institute from 1911 in Green Harbour, from 1916 in Longyearbyen and from 1950 in Ny-Ålesund (Førland et al. 1997b). Glacier mass balance measurements started 1966 at Brøggerbreen (J. O. Hagen, pers. comm 2002), while the fi rst known systematic runoff measure- ments in Svalbard were started in Bayelva close to Ny-Ålesund from 1974 (Repp 1988a, b). Reg- ular and continuous runoff measurements have been operated by the Norwegian Water Resourc- es and Energy Directorate since 1989 in Bay elva. Later, regular runoff measurements have been established in three other catchments: Endalen/ Isdammen and De Geerdalen close to Longyear- byen, and in Londonelva near Ny-Ålesund. Evap- oration measurements are almost non-existent, but a few measurements were done by the Foun- dation for Technical and Industrial Research at the Norwegian Institute of Technology (SINTEF) in Ny-Ålesund in 1992 and 1993 as part of the Land Arctic Physical Processes (LAPP) project (Institute of Hydrology et al. 1999). The fi rst known water balance Study in Sval- bard was done near Kings Bay in 1968 (Geoffray 1968). Data from Repp’s studies in Bayelva were used by Bruland (1991) in an attempt to estab- lish the water balance and calibrate hydrological models in Svalbard. These data were also used by Hagen & Lefauconnier (1993). Some other river fl ow measurements have been reported by Rus- sian scientists (Gokhman & Khodakov 1986) and Polish scientists (Pulina et al. 1984). None of these studies included a complete water balance computation. Jania & Pulina (1994) reported on the results from several Polish research projects in Sval- bard, and summarized the results in tables for water balance and chemical denudation in the Weren skioldbreen basin for one hydrological year (1979/80) and for several unidentifi ed basins in the Horn sund area. The main results are sum- marized in Table 1. Water balance computation for catchments in Svalbard based on Norwegian studies was pre- sented at the Northern Research Basins meeting in Ny-Ålesund in 1994 (Killingtveit et al. 1994). This study was based on the observations initi- ated by the Norwegian National Committee for Hydrology from 1991 to 1993 in the three catch- ments—Bayelva, De Geerdalen and Endalen/ Isdammen. In this paper the results from a litera- ture search were also reported, but no other stud- ies beyond those already noted could be found. Several studies have been done concerning winter water balance and possible groundwa- ter fl ow to Tvillingvatnet, near Ny-Ålesund, and Isdammen, near Longyearbyen. These are sum- marized in the section on groundwater runoff. Sand & Bruland (1999) presented the water bal- ance for the Bayelva, Londonelva and De Geerel- Table 1. Water balance results from Polish research projects in Svalbard reported by Jania & Pulina (1994). Glaciated catchments Permafrost basins PA 1070 - 1340 mm/year 800 mm/year QS 1680 - 1920 mm/year 710 mm/year EA 90 mm/year 90 mm/year ∆MG 670 - 700 mm/year 0 mm/year Table 2. Water balance results from catchments in Spits- bergen reported by Sand & Bruland (1999). Bayelva Londonelva De Geerdalen PA (mm/year) 968 595 478 EA (mm/year) 46 100 108 ∆MG (mm/year) –456 0 –21 QS (mm/year) 1053 411 525 ε (mm/year) –7 118 –134 167Killingtveit et al. 2003: Polar Research 22(2), 161–174 va catchments for the years 1991 to 1998 (see Table 2). For years when snow survey data existed they determined the areal winter precipitation based on the snow survey data rather than measurement of precipitation. For other years, the areal precipi- tation was estimated from precipitation measure- ments only. A recent work (Mercier 2001) includes a sum- mary from a large number of studies of geo- morphology, glaciology and hydrology in Sval- bard, including studies of the components in the hydrological balance (water balance) for Austre Lovénbreen in Kongsfjorden and for a 0.1 km2 non-glaciated area near Zeppelinerfjellet near Ny-Ålesund. Some of the results from this report are included in subsequent sections. Catchments selected for water balance study In this paper, data from the three catchments with the best data coverage—Bayelva, De Geerdalen and Endalen—will be used. A brief description of the different observation data (runoff, precipita- tion, evaporation and glacier balance) is given in the following sections. The location of the catch- ments and a simplifi ed map of each catchment is shown in Fig. 2. The area–elevation distribution within each catchment (hypsographic curve) is shown in Fig. 3. A summary of the most impor- tant data for each catchment is given in Table 3. Runoff measurements–stations and records Surface runoff—Bayelva. Runoff measure- ments in Bayelva started in 1974 (Repp 1979). These measurements ended in 1978, but were resumed in 1990 as part of an initiative taken by the Norwegian National Hydrological Commit- tee in 1987. Measurements during the fi rst year are considered unreliable due to an unstable river bed. In 1988 the river profi le was stabilized by a concrete Crump-type weir and a permanent and stable rating curve could be established (Skrette- berg 1992). The main operational problem now is caused by snow blocking the weir at the start of the snowmelt season, leading to errors in calcula- tion of river discharge. This is corrected for as far as possible, but the resulting accuracy is not pre- cisely known. Surface runoff—De Geerdalen. In the autumn of 1990 a second runoff station was established in De Geerdalen. The station is located in a narrow gorge with stable bedrock profi le close to the outlet of the river, at Hyperittfossen (Fig. 2). The station has been in routine operation since summer 1991. The main operational problems here are related to ice and snow blocking in the narrow gorge. Surface runoff—Isdammen/Endalen. The fi rst station was established early in the summer of 1992, in the outlet from Isdammen, the water supply reservoir for Longyearbyen (Fig. 2). Here, the main operational problems were caused by uncertain rating curves and ungauged leakage in the outlet structure. Later, the station was re- established in Endalselva, upstream of a road cul- vert close to the inlet to Isdammen. At this site bottom ice build-up in the river profi le seems to be the main operational problem, possibly also some leakage outside the main culvert. There is some uncertainty in the drainage area for these catchments, since the glacier Bogerbreen may also have some drainage to another catchment. Annual runoff for hydrological years for the three catchments is shown in Table 4. Average monthly runoff (mm/month) is shown in Table 5 and the seasonal distribution of the runoff over the year is shown in Fig. 4. It is evident from Fig. 4 that runoff occurs mainly in June–September, with maximum runoff occurring in July. The runoff is dominated by snowmelt in June and July, while in August and September the runoff mainly comes from rainfall and glacial melt. The high percentage of glaciers in the Bayelva catch- ment gives a relatively higher runoff in August and September than in the two other catchments. Groundwater runoff. Due to the permafrost there is usually no fl ow in groundwater below the Table 3. Data for the three catchments Bayelva, De Geerdalen and Isdammen/Endalen. The minimum elevation is where the runoff gauging station is located Area (km2) Elevation (m a.s.l.) % glaciersMin. Average Max. Bayelva 30.9 4 265 742 55 De Geerdalen 79.1 40 410 987 10 Isdammen 34.4 3 427 1015 17 Endalselva 28.8 4 427 1015 20 168 Water balance investigations in Svalbard active zone (1 - 2 m) and therefore probably no groundwater fl ow directly to the sea. Exceptions may be groundwater fl ow from recharge areas under temperate glaciers and discharge through taliks under lakes or directly to the sea. Speculations concerning possible ground- water infl ow to the water supply reservoirs for Ny-Ålesund and Longyearbyen (Winther 1994) spurred special studies of the winter water bal- ance for these reservoirs and their catchments. For Tvillingvatnet (Ny-Ålesund) there was pre- viously documented winter infl ow from ground- water in the 1920s. This groundwater fl ow was probably generated by recharge below the tem- perate basal part of the glacier Austre Brøgger- breen (Haldorsen & Heim 1999). Recent studies during 1990s (Sandsbråten 1995) did not confi rm these results, but fl ow may have been reduced due to reduced infi ltration area as the glacier has retreated signifi cantly (Haldorsen & Heim 1999). Groundwater recharge in the Isdammen reser- voir near Longyearbyen have been postulated as a possibility. Studies during the winter 1999/ 00 have not lead to conclusive evidence of any groundwater infl ow (Klungland 2000). There are no data to confi rm or exclude the possibility of groundwater fl ow directly to the sea in Svalbard. Halvorsen & Heim (1999) did a detailed investigation along Kongsfjorden in the Ny-Ålesund area but could not fi nd any indica- tions of springs discharging along the coast or into the fjord. Therefore, in the water balance cal- culation presented here, it is assumed that no sig- nifi cant groundwater fl ow exists into or from the three catchments studied. Fig. 4. Seasonal variations in runoff for the three selected catchments. Monthly averages computed for the observation period. Table 5. Seasonal variation in runoff, mm/month, average for years 1990–2001. Runoff (mm) Bayelva De Geerdalen Endalen January 0 0 0 February 0 0 0 March 0 0 0 April 0 0 0 May 0 1 0 June 201 164 109 July 437 233 247 August 317 106 114 September 111 33 34 October 4 4 0 November 0 0 0 December 0 0 0 Table 4. Annual runoff (mm/year) for the three catchments. Hydrol. year Bayelva De Geerdalen Endalen 1990 1288 573 1991 947 489 1992 1097 641 486 1993 1292 429 596 1994 962 481 1995 1005 463 429 1996 1012 596 555 1997 1008 605 608 1998 1061 472 611 1999 1227 593 534 2000 877 586 2001 1316 573 Average 1091 539 545 169Killingtveit et al. 2003: Polar Research 22(2), 161–174 Precipitation measurements Stations and data records Bayelva. There are no regular precipitation meas- urements in the Bayelva catchment, but the mete- orological station in Ny-Ålesund is located quite close to the catchment. Here, regular precipita- tion measurements have been made since 1950. In addition, various studies concerning precipi- tation correction and precipitation distribution have been done. One of the most interesting is a study of precipitation distribution and precip- itation gradients carried out by the Norwegian Meteorological Institute (Førland et al. 1997a). In other studies precipitation has also been estimat- ed indirectly through snow measurements and glacier mass balance measurements within the catchment. These measurements are later used to establish precipitation–elevation gradients. De Geerdalen. There are no regular precipita- tion measurements in the catchment and the clos- est meteorological station is located at Svalbard Airport, about 20 km south-west of the valley De Geerdalen. A number of precipitation stations were operated for a few years in the early 1990s as part of the fi rst water balance studies initiated by the Norwegian Hydrological Committee (Kill- ingtveit et al. 1994). A network of snow measure- ment stations has also been operated since 1991, making it possible to estimate winter precipita- tion indirectly (Tveit & Killingtveit 1994). These measurements have been used to estimate pre- cipitation–elevation gradients in the catchments (Killingtveit et al. 1994). Isdammen/Endalen. There are no regular precip- itation measurements in the catchment and the closest meteorological station is located at Sval- bard Airport, about 10 km west of Isdammen. A number of precipitation stations were operat- ed a few years in the early 1990s as part of the fi rst water balance studies. These measurements have been used to estimate precipitation–eleva- tion gradients in the catchments (Killingtveit et al. 1994). Annual measured precipitation in Ny-Ålesund and Svalbard Airport is shown for the hydrologi- cal years 1990/91 to 2000/01 in Table 6. Average (normal) monthly precipitation for the two sta- tions is shown in Table 7. Areal precipitation Precipitation correction factors based on previ- ous studies from Svalbard (Repp 1979; Hagen & Lefauconnier 1993, 1996; Tveit & Killingtveit 1994; Hanssen-Bauer et al. 1996; Sand et al. 2003) and other relevant studies in other Arctic catchments were used in this water balance cal- culation, leading to an average correction of 1.15 for rainfall and 1.65 for snow precipitation for Ny-Ålesund data and slightly higher values (1.15 and 1.75) for Svalbard Airport data. In these water balance computations a gradient of 15 %/ 100 m was selected for the Isdammen/Endalen and Bayelva catchments, and 20 %/100 m for De Table 6. Annual precipitation data for stations at Ny-Ålesund and Svalbard Airport (mm/year). Calender year Ny-Ålesund Svalbard Airport 1990 479 157 1991 502 257 1992 381 194 1993 674 262 1994 385 220 1995 256 159 1996 547 234 1997 390 217 1998 248 92 1999 348 187 2000 519 203 2001 487 185 Average 431 196 Table 7. Precipitations averages (mm/month) for the period 1961–1990 based on data from stations at Ny-Ålesund and Svalbard Airport. Ny-Ålesund Svalbard Airport January 32 15 February 36 19 March 45 23 April 23 11 May 18 6 June 18 10 July 28 18 August 38 23 September 46 20 October 37 14 November 33 15 December 31 16 Sum 385 190 170 Water balance investigations in Svalbard Geerdalen. Observed precipitation was fi rst corrected for catch errors and then for elevation gradients, giving the average (areal) precipitation for each of the catchments. For catch errors the rain cor- rection factor was applied for months with air temperature > 0 °C and snow correction factor for months with air temperature < 0 °C. The results of these computations are given as annual values in Table 8 for the Bayelva catchment. The same computations were done for the two other catch- ments. This is the fi nal and total precipitation input to the three catchments and it is used in the subsequent water balance computation. Evaporation measurements Evaporation measurements have been and remain very scarce in Svalbard. A survey by Bruland (1991) found only one reference to previous meas- urements in Svalbard. Based on this observation and a few data from other Arctic sites, Bruland estimated an actual average annual evapora- tion of 100 mm/year. This value was later also used by Hagen & Lefauconnier (1993, 1996). In the fi nal report from the LAPP project (Institute of Hydrology et al. 1999) the use of evaporation data in water balance computations is discussed with reference to Hagen & Lefauconnier (1993) and Killingtveit et al. (1994). The value of 100 mm/year is used, the report stating that “Still, no better estimate exists, and we have used the same value in our study” (p. 38). Net evaporation from glaciers was assumed to be 0. Jania & Pulina (1994) used 90 mm/year in their water balance calculations. This value seems to be based on calculations from meteorological data “due to diffi culties in taking measurements (a large error range is likely)” (p. 62). To improve evaporation estimates, a Class A Evaporation pan was installed in Ny-Ålesund in 1992 and operated by SINTEF. Some early results from the measurements are reported in Killingtveit et al. (1994). The annual pan evap- oration was estimated to 166 mm, based on pan measurements and some computed values where a regression equation between evaporation and air temperature was used to infi ll data gaps. The average actual evaporation from non-glaciated catchments was estimated to 120 mm/year in the Ny-Ålesund area, at an elevation of 10 m a.s.l. Mercier (2001) reports computed potential evap- oration for the years 1969–1995, based on both the Turc and the Penman formulas. Average com- puted values were 200 mm/year (Penman) and 51 mm/year (Turc). The large deviation between the two methods is discussed and it seems that most confi dence is placed on the results from the Turc formula. In this study a computation of potential evapo- ration as a function of air temperature was used. This function was calibrated by regression anal- ysis using SINTEF’s evaporation observations in Ny-Ålesund and air temperature data from Ny- Ålesund. The computed average annual poten- tial evaporation for the 12-year period 1989–1991 was 138 mm/year for an area close to sea level in Ny Ålesund, computed from air temperature data and assuming a Pan coeffi cient of 1. Evapo- ration from snow or glaciers is not included. This value can be compared to data from Axel Heib- erg Island (80° N) where Ohmura (1982) found an average evaporation of 138 mm/year. This value also included an estimated value of 20 mm/year of sublimation from snow. Using the regression model and observed air temperature data the average evaporation from glacier-free areas could be computed for each year, using an average lapse rate of –0.6 °C/ 100 m. The results were 80 mm/year in Bayelva and 82 mm/year in De Geerdalen and Endalen— about 50 mm/year less than potential evaporation close to sea level. Net annual evaporation from glaciers is assumed to be 0, as in previous studies in Sval- bard (Hagen & Lefauconnier 1993, 1996; Insti- Table 8. Areal precipitation calculation in Bayelva (mm/ year). Hydrological year Observed precip. Catch correction Elevation correction Areal precip. 1990/91 472 252 293 1016 1991/92 420 182 243 845 1992/93 396 213 247 856 1993/94 678 369 424 1472 1994/95 169 84 102 355 1995/96 648 348 403 1398 1996/97 312 157 190 659 1997/98 242 148 158 548 1998/99 380 175 225 780 1999/00 387 148 217 751 2000/01 582 207 320 1109 Average 426 207 256 890 171Killingtveit et al. 2003: Polar Research 22(2), 161–174 tute of Hydrology et al. 1999). This assumption can be questioned, but it has not been possible to fi nd better estimates from studies in Svalbard. There will probably be some losses by sublima- tion from dry snow and blowing snow but also some input by sublimation on ice during summer months. The net balance between losses and gains by sublimation on glaciers in Svalbard is probably small but cannot be verifi ed yet. More studies need to be done. If such errors exist, posi- tive or negative, it will add up in the error term in the water balance. Using these data and the percent of glaciat- ed area for each catchment (Table 3), the follow- ing values for average annual actual evaporation were computed for the catchments: Bayelva, catchment average 80 mm/year * 0.45 = 38 mm/year at 265 m a.s.l. De Geerdalen, catchment average 82 mm/year * 0.90 = 72 mm/year at 410 m a.s.l. Endalen, catchment average 82 mm/year * 0.80 = 66 mm/year at 427 m a.s.l. Storage terms If the water balance is computed on an annual basis and for hydrological years, most of the storage terms in Eq. (2) can be neglected. Using a hydrological year from 1/10 to 30/9, we can assume that changes in snow storage, lakes, ground water and soil moisture from one year to another can be neglected. Lake storage is only important in Endalen/ Isdammen but here the lake level has been meas- ured and the lake is always fi lled to the same (maximum) level at the beginning of the winter season. Due to the deep permafrost there is no groundwater recharge from the surface except possibly below some of the glaciers. Such ground- water recharge has been not been identifi ed in these three catchments, though there may be some small recharge under Vestre Lovénbreen. (Haldorsen & Heim 1999). If any recharge exists it is probably very small; there are no indications of substantial changes from year to year. There are, to our knowledge, no indications that water content in the active layer changes signifi cantly from year to year. However, no measurements are available to prove this assumption. Theoretical- ly, a very dry summer and autumn could dry out the soil moisture in the active layer, leaving a def- icit that would fi ll up next spring. Also, there are no wetlands or icings that could store and “carry over” water from one year to another. Seasonal snow cover is always melted before the end of September and new snow cover usually starts to build up from October. Therefore, it seems acceptable to assume that all these storage terms can be neglected, except possibly some chang- es in soil moisture in the active layer. Since we have no data to compute such changes, the possi- ble errors will contribute to the error term in the water balance. The change in glacier mass bal- ance from year to year is very important and must to be included in the water balance. Mean annual net mass balances for a number of Spitsbergen glaciers have been studied for many years, some back to 1951. At Austre Brøgger- breen, the annual mass balance has been studied since 1967, at Longyearbreen from 1977 to 1982 Table 9. Glacier mass balance for Austre Brøggerbreen (in m) for hydrological years 1974/75 to 2000/01. Hydrological year Winter Summer Total 1974/75 0.78 –1.09 –0.31 1975/76 0.72 –1.17 –0.45 1976/77 0.76 –0.87 –0.11 1977/78 0.75 –1.31 –0.56 1978/79 0.77 –1.48 –0.71 1979/80 0.75 –1.27 –0.52 1980/81 0.46 –1.01 –0.55 1981/82 0.64 –0.68 –0.04 1982/83 0.70 –0.97 –0.27 1983/84 0.69 –1.42 –0.73 1984/85 0.93 –1.48 –0.55 1985/86 0.98 –1.3 –0.32 1986/87 0.82 0.6 0.22 1987/88 0.61 –1.13 –0.52 1988/89 0.56 –1.01 –0.45 1989/90 0.75 –1.41 –0.66 1990/91 0.92 0.79 0.13 1991/92 0.69 –0.89 –0.1 1992/93 0.54 –1.57 –1.03 1993/94 0.79 –0.95 –0.16 1994/95 0.56 –1.34 –0.78 1995/96 0.78 –0.95 –0.17 1996/97 0.50 –1.12 –0.88 1997/98 0.65 –1.78 –1.13 1998/99 0.51 –0.87 –0.36 1999/00 0.41 –0.51 –0.11 2000/01 – – –0.45 172 Water balance investigations in Svalbard and at Bogerbreen from 1975 to 1986 (J. Kohler, pers. comm. 2002). Annual results are given in Table 9. The average annual mean net balance during the observational period is given below: Brøggerbreen 1967–2002, –450 mm/year Bogerbreen 1975–1986, –430 mm/year Longyearbreen 1977–1982, –550 mm/year For water balance computations in Bayelva we used the changes measured at Brøggerbreen directly. For the water balance computations in the De Geerdalen and in Endalen/Isdammen there exist no glacier mass balance data for the glaciers in the catchments for the period when runoff has been measured. Since the Bogerbreen and Longyearbreen glaciers are much closer to these catchments than Brøggerbreen, we assumed that the Bogerbreen and Longyearbreen glaciers provide better information on glacial mass bal- ance than Brøggerbreen. In lieu of direct glacier mass balance data in De Geerdalen and Endalen/ Isdammen, we used annual changes measured at Brøggerbreen as an index of glacier mass bal- ance, scaled by the difference between the gla- ciers Bogerbreen and Longyearbreen and Brøg- gerbreen during the overlapping measurement periods. The computed scaling factor is 1.06, i.e. changes in glacier storage in De Geerdalen and Endalen are computed as 1.06 times changes at Brøggerbreen, per unit area. Water balance calculation The water balance for Bayelva was calculated annually from data given in Tables 4, 8 and 9. The results are summarized in Table 10. The water balance for De Geerdalen and Endalen/Isdammen was calculated similarly and the results are summarized as average for the whole computational period in Table 11, together with the results from Bayelva. Discussion of results The average water balance is good for all three catchments, with an average error term (imbal- ance) close to 0. There are, however, large devi- ations in some years which cannot easily be explained. A correlation analysis shows that there is still a strong correlation between the annual error term and winter precipitation. This indi- cates that the residual errors are probably relat- ed to problems of precipitation correction and areal precipitation computations. The results show very clearly the large difference between observed precipitation and areal precipitation computed within the catchments. The observed precipitation has to be multiplied by a factor on the order of 2 to compensate for errors in meas- urements and non-representative locations. The glacial mass balance term also makes a large con- tribution to the water balance and possible errors here will easily contribute signifi cantly to the errors in some years. The evaporation is usually the most uncertain term in water balance compu- tations. The results here cannot verify or disprove the evaporation estimates since the error terms in many years is on the same order of magnitude as the evaporation. Comparing our results with the results report- ed by Jania & Pulina (1994), both the measured runoff and precipitation estimates are considera- Table 10. Water balance for the De Geerdalen catchment in hydrological years 1990/91 to 2000/01: Pwinter + Psummer + ∆glaciers – Q – E = ε. All terms are in mm. Hydrological year Pwin Psum ∆glaciers Q E ε 1990/91 504 176 -14 573 70 23 1991/92 329 270 11 489 69 52 1992/93 503 134 109 641 93 12 1993/94 455 254 17 429 45 252 1994/95 224 115 83 481 77 –136 1995/96 561 189 18 463 53 252 1996/97 347 197 93 596 50 –8 1997/98 254 68 120 605 103 –266 1998/99 218 232 38 472 69 –54 1999/00 316 208 12 593 65 –122 2000/01 288 180 48 586 95 –166 Average 364 184 49 539 72 –15 Std. dev. 122 60 45 72 19 162 Table 11. Water balance for all catchments – Annual average for hydrological year (Pwinter + Psummer + ∆Glaciers – Q – E = ε. All terms are in mm. Catchment Pwin Psum ∆Glaciers Q E ε Std ε Bayelva 597 277 245 1050 37 31 230 De Geerdalen 364 184 49 539 72 -15 162 Endalen/ Isdammen 321 158 101 545 66 -14 106 173Killingtveit et al. 2003: Polar Research 22(2), 161–174 bly lower in this study. It should be kept in mind the catchments investigated by Jania & Pulina (1994) are all located in the Hornsund area, in southern Spitsbergen, while our catchments are located in the central and north-western parts of the island. Sand et al. (2003) did a study of region- al snow distribution on Spitsbergen which indi- cates that the Hornsund area receives approxi- mately twice as much precipitation during the snow accumulation period (October–May) as the central region. So far, the results from the water balance stud- ies carried out in Svalbard have not been evalu- ated against water balance studies from other regions of the Arctic. However, the results from Svalbard will be part of a recently initiated study which will be an intercomparison of water bal- ance in Arctic experimental watersheds (D. L. Kane, pers. comm 2003). Further studies—recommendations The results show that the calculated water bal- ance in Svalbard still cannot be considered good enough to assess the individual components of the hydrological cycle with fair accuracy. Even though the average balance (error term) in the three catchments is close to zero, errors in indi- vidual years are still considerable. This indicates that it is still necessary to improve data collection and possibly also correction methods, in particu- lar for precipitation and glacial net balance. We recommend more detailed investigations concerning precipitation distribution, in partic- ular the distribution during winter, when most of the precipitation falls. Snow measurements can be used to study amount and distribution of seasonal snow. Such data can be useful both to improve methods for precipitation correction and methods for computing areal precipitation in Svalbard. Even if evaporation is not a dominant factor in the water balance, it would be very useful to improve the data base and we recommend that regular measurements should be established to acquire more precise data and to study seasonal and interannual variations in evaporation. Also, it would be useful to study and collect data on the evaporation from snow since this has not yet been studied in Svalbard. Acknowledgements.—We would like to thank all those who have supplied data and information concerning previous stud- ies and reports. In particular we thank the Norwegian Mete- orological Institute for supplying all meteorological data, and the Norwegian Water Resources and Energy Directorate for supplying runoff data and information concerning quality of runoff data. We also thank to Jack Kohler at the Norwegian Polar Institute and Jon Ove Hagen at the University in Oslo for providing information about glacier mass balance for a number of glaciers. References Bogdanova, E. G. 1968: Estimate of the reliability of the char- acteristics of the shortage in solid precipitation due to wind. Soviet Hydrology. Selected Papers 2 1968, 139–146. Wash- ington, D. C.: American Geophysical Union. Bruland, O. 1991: Vassbalanse og avlaupsmodellar i perma- frostområder. (Water balance and runoff in permafrost areas). Thesis D-1991-28. Dept of Hydraulic and Environ- mental Engineering, University of Trondheim. Førland, E. J., Allerup, P., Dahlström, B., Elomaa, E., Jóns- son, T., Madsen, H., Perälä, J., Rissanen, P., Vedin, H. & Vejen, F. 1996: Manual for operational correction of Nordic precipitation data. Rep. 24/96. Oslo: Norwegian Meteorological Institute. Førland, E., Hanssen-Bauer, I. & Nordli, P. Ø. 1997a: Oro- graphic precipitation at the glacier Austre Brøggerbreen, Svalbard. Klima 02/97. Oslo: Norwegian Meteorological Institute. Førland, E., Hanssen-Bauer, I. & Nordli, P. Ø. 1997b: Cli- mate statistics & longterm series of temperature and pre- cipitation at Svalbard and Jan Mayen. Klima 21/97. Oslo: Norwegian Meteorological Institute. Geoffray, H. 1968: Etude du bilan hydrologique et de l’érosion sur un bassin pertiellement englacé, Spitsberg, Baie du Roi, 79° Lat. Nord. (A study of water balance and erosion in a partly glaciated basin, Spitsbergen, Kings Bay, 79° N.). Thesis, University of Rennes. Gokhman, V. V. & Khodakov, V. G. 1986: Hydrological investigations in the Mimer river basin, Svalbard in 1983. Polar Geogr. Geol. 10, 309–316. Hagen, J. O. & Lefauconnier, B. 1993: Reconstructed runoff from the High Arctic basin Bayelva in Svalbard based on mass-balance measurements. In K. Sand (ed.): Polar hydrologi. Rapport fra forskermøte i Trondheim 29–30 mars 1993. (Report from research meeting in Trondheim 29–30 March 1993.) SINTEF Rep. STF60 A93081. Pp. 25– 38. Trond heim: Norwegian Institute of Technology. Hagen, J. O. & Lefauconnier, B. 1996: Reconstructed runoff from the High Arctic Basin Bayelva based on mass-balance measurements. Nord. Hydrol. 26, 285–296. Haldorsen, S. & Heim, M. 1999: An Arctic groundwater system and its dependence upon climate change: an exam- ple from Svalbard. Permafrost Periglacial Process. 10, 137–149. Hanssen-Bauer, I., Førland, E. & Nordli, P. Ø. 1996: Meas- ured and true precipitation at Svalbard. Klima 31/96. Oslo: Norwegian Meteorological Institute. Institute of Hydrology, Institute of Terrestrial Ecology, Finn- ish Meteorological Institute, Finnish Environmental Insti- tute, Foundation for Technical and Industrial Research 174 Water balance investigations in Svalbard (Norwegian Institute of Technology), Institute of Geogra- phy & University of Copenhagen 1999: Final report LAPP: Land Arctic Physical Processes. Available on the internet at http://www.nwl.ac.uk/ih/www/research/iresearch.html. Jania, J. & Pulina, M. 1994: Polish hydrological studies in Spitsbergen, Svalbard: a review of some results. In K. Sand & Å. Killingtveit (eds.): Proceedings of the 10th Interna- tional Northern Research Basins Symposium and Work- shop, Spitsbergen, Norway. Pp. 47–76. SINTEF Rep. 22 A96415. Trondheim: Norwegian Institute of Technology. Killingtveit, Å., Petterson, L.-E. & Sand, K. 1994: Water bal- ance studies at Spitsbergen, Svalbard. In K. Sand & Å. Killingtveit (eds.): Proceedings of the 10th Internation- al Northern Research Basins Symposium and Workshop, Spitsbergen, Norway. SINTEF Rep. 22 A96415. Pp. 77–94. Trondheim: Norwegian Institute of Technology. Klungland, K. O. 2000: Winter water balance for Isdammen, Svalbard. Thesis, Stavanger University College and Uni- versity Centre on Svalbard. Larsson, L. W. & Peck, E. L. 1974: Accuracy of precipitation measurements for hydrological modelling. Water Resour. Res. 10, 857–863. Mercier, D. 2001: Le Ruissellement au Spitsberg. (Runoff on Spitsbergen.) Clermont-Ferrand, France: Blaise Pascal University Press. Ohmura, A. 1982: Evaporation from the surface of the arctic tundra on Axel Heiberg Island. Water Resour. Res. 18, 291– 300. Petterson, L.-E. 1994: The hydrological regime of Spitsber- gen, Svalbard. In K. Sand & Å. Killingtveit (eds.): Pro- ceedings of the 10th International Northern Research Basins Symposium and Workshop, Spitsbergen, Norway. SINTEF Rep. 22 A96415. Pp. 95–107. Trondheim: Norwe- gian Institute of Technology. Pulina, M., Rereyma, J., Kida, J. & Krewczyk, W. 1984: Characteristics of the polar hydrological year 1979/80 in the basin of the Werenskiold glacier, SW Spitsbergen. Pol. Polar Res. 5, 165–182. Repp, K. 1979: Breerosjon, glasiohydrologi og materialtran- sport I et høyarktisk miljø, Brøggerbreene, Vest-Spitsber- gen. (Glacial erosion, glacial hydrology and sediment transport in a High Arctic basin, Brøggerbreen, west Spits- bergen.) MSc thesis, University of Oslo. Repp, K. 1988a: The hydrology of Bayelva, northwest Spits- bergen. In T. Thomsen et al. (eds.): Proceedings of the 7th Northern Research Basins Symposium/Workshop. May 25- June 1 1988. Illulissat, Greenland. Pp. 105–114. Copenha- gen: Danish Scoeity for Arctic Technology. Repp, K. 1988b: The hydrology of Bayelva, Spitsbergen. Nord. Hydrol. 4, 259–268. Sand, K. & Bruland, O. 1999: Water balance of three High Arctic river basins in Svalbard. In J. Elíasson (ed.): Pro- ceedings of the 12th International Northern Research Basins Symposium and Workshop. Reykjavik, Kirkjubæjar- klaustur and Höfn, Iceland, August 23–27, 1999. Pp. 270– 283. Reykjavik: Engineering Institute, University of Ice- land. Sand, K., Winther, J.-G., Maréchal, D., Bruland, O. & Mel- vold, K. 2003: Regional variations of snow accumulation on Spitsbergen, Svalbard, 1997–99. Nord. Hydrol. 34, 17– 32. Sandsbråten, K. 1995: Vannbalanse i et lite arktisk nedbør- felt, Tvillingvatn, Svalbard. (Water balance in a small Arctic catchment, Tvillingvann, Svalbard.) Repo. 43. Dept. of Geography, University of Oslo. Sevruk, B. 1982: Methods of correction for systematic error in point precipitation measurement for operational use. WMO Oper. Hydrol. Rep. 21. WMO-NO 589. Geneva: World Meteorological Organization. Skretteberg, R. 1992: The establishment of gauging stations under Arctic conditions—the Svalbard experience. In T. D. Prowse et al. (eds.): Proceedings of the 9th Internation- al Northern Research Basins Symposium and Workshop, Canada 1992. NHRI symposium no. 10, vol. 2. Pp. 509–518. Saskatoon: National Hydrology Research Institute. Strutzer, L. R. 1965: Principal shortcomings of methods of measuring atmospheric precipitation and means of improv- ing them. Soviet Hydrology, Selected Papers, 1 1965, 21– 35. American Geophysical Union. Tveit, J. & Killingtveit, Å. 1994: Snow surveys for studies of water budget on Svalbard 1991–1994. In K. Sand & Å. Killingtveit (eds.): Proceedings of the 10th Internation- al Northern Research Basins Symposium and Workshop, Spitsbergen, Norway. SINTEF Rep. 22 A96415. Pp. 489– 509. Trondheim: Norwegian Institute of Technology. Winther, J.-G. 1994: Polar hydrology—Svalbard. Revision of the original R&D programme. In K. Sand & Å. Killingtveit (eds.): Proceedings of the 10th International Northern Research Basins Symposium and Workshop, Spitsbergen, Norway. SINTEF Rep. 22 A96415. Pp. 1–22. Trondheim: Norwegian Institute of Technology. Winther, J.-G., Bruland, O., Sand, K., Killingtveit, Å. & Marechal, D. 1998: Snow accumulation distribution on Spitsbergen, Svalbard in 1997. Polar Res. 17, 155–164. << /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles true /AutoRotatePages /All /Binding /Left /CalGrayProfile (Dot Gain 20%) /CalRGBProfile (sRGB IEC61966-2.1) /CalCMYKProfile (U.S. Web Coated \050SWOP\051 v2) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Warning /CompatibilityLevel 1.4 /CompressObjects /Tags /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages true /CreateJDFFile false /CreateJobTicket false /DefaultRenderingIntent /Default /DetectBlends true /ColorConversionStrategy /LeaveColorUnchanged /DoThumbnails false /EmbedAllFonts true /EmbedJobOptions true /DSCReportingLevel 0 /SyntheticBoldness 1.00 /EmitDSCWarnings false /EndPage -1 /ImageMemory 1048576 /LockDistillerParams false /MaxSubsetPct 100 /Optimize true /OPM 1 /ParseDSCComments true /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveEPSInfo true /PreserveHalftoneInfo false /PreserveOPIComments false /PreserveOverprintSettings true /StartPage 1 /SubsetFonts true /TransferFunctionInfo /Apply /UCRandBGInfo /Preserve /UsePrologue false /ColorSettingsFile () /AlwaysEmbed [ true ] /NeverEmbed [ true ] /AntiAliasColorImages false /DownsampleColorImages true /ColorImageDownsampleType /Bicubic /ColorImageResolution 300 /ColorImageDepth -1 /ColorImageDownsampleThreshold 1.50000 /EncodeColorImages true /ColorImageFilter /DCTEncode /AutoFilterColorImages true /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /ColorImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000ColorACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000ColorImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasGrayImages false /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 300 /GrayImageDepth -1 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages true /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /GrayImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000GrayACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000GrayImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasMonoImages false /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 1200 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict << /K -1 >> /AllowPSXObjects false /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile () /PDFXOutputCondition () /PDFXRegistryName (http://www.color.org) /PDFXTrapped /Unknown /Description << /FRA /ENU (Use these settings to create PDF documents with higher image resolution for improved printing quality. The PDF documents can be opened with Acrobat and Reader 5.0 and later.) /JPN /DEU /PTB /DAN /NLD /ESP /SUO /ITA /NOR /SVE >> >> setdistillerparams << /HWResolution [2400 2400] /PageSize [612.000 792.000] >> setpagedevice