User manual of the (KNMI North Sea Wind) KNW-atlas
Contents
1. direction sector so 10 means 1 In the last column we see the frequency of occurrence of wind speeds irrespective of wind direction per bin in percent so not in 0 1 percent Figure 10 shows that according to the KNW atlas the prevailing wind direction at MMIJ at 100 m height is SSW 195 225 degrees because the highest figure in the fourth row ff gt 0 is 15 2 Similarly the prevailing wind speed is 9 10 m s 5 Beaufort because 8 1 is the highest figure in the column on the far right So you can use the table to find the prevailing wind direction or the prevailing wind speed but you cannot use the table to find the most prevailing combination of wind direction and wind speed It would appear to be NNE 15 45 degrees and 8 9 m s between 4 and 5 Beaufort because the highest number in the lower part of the table is 104 10 4 but this interpretation is incorrect This is because the percentage in the bin is relative to the total number of data for the wind direction sector NNE and the wind blows from these directions much less often than from SSW The highest value in the SSW column 7 5 represents more data than the 10 4 of data from NNE One glance at the wind rose of figure 12 makes this clear where the percentages are relative to all the data and not just to those of a given wind direction sector The fact that 10 4 is the highest value is also partly because NNE has the lowest average wind speed so the occurrences are for2. 02 10 2014 01 01 Time wind speed m s 1 Figure 14 An example of a grid cell time series nearest grid cell to MMIJ showing the hourly average wind speed 4 4 KNW KNMI North Sea Wind Atlas grid cell time series ASCII CSV The time series that can be downloaded here contain the same information as those of the previous section The main difference is the format of the files and that there are individual files for each weather parameter and height combination ASCII CSV files do not require separate sub datasets with structural information concerning time and locations 5 How to use and not to use the KNW atlas The wind speed of the KNW atlas has been validated against measured averages ranging from one year OWEZ to ten years QuikScat Based on this validation the conclusion can be drawn that the quality of the KNW atlas is high at these time scales less than 0 5 m s at 10 m and less than 0 2 m s at greater heights and moreover comparable to climatology based on measurements for heights above 10 m 5 1 Diurnal cycle The KNW atlas has not been thoroughly validated on shorter timescales months days or hours However figure 15 shows a comparison with measurements for two periods of 10 days with both extremely high and low winds The KNW and measured time series match quite well So even on the time scale of 1 hour the KNW 19 wind speed can be used effectively The high winds from 10 20 January are from different di
3. 10 year extreme wind speeds for heights around wind turbine hub height figure 2 Also a first attempt was made to validate the 10 m wind speeds of the KNW atlas against scatterometer winds 10 m wind speeds derived from satellite measurements Compared to the QuikSCAT 25 km product 1999 2009 the KNW atlas overestimates the 10 m wind speed for most of the North Sea by less than 0 5 m s figure 3 This is the opposite for the southern part of the North Sea including the areas Borssele and Hollandse Kust designated for wind energy exploitation Here the KNW atlas underestimates the 10 m wind speed by less than 0 5 m s Comparing these results to the vertical validation results shows that the validation at 10 m should not be used to estimate the validity of the KNW winds at the heights relevant for wind energy purposes _ Wind speed versus height FINO 1 HARMONIE rr FINO 1 SONIC o o w o Wind speed m s o rT Try LIDA 8 w T PTF Tr Height m Wind speed versus height MMIJ HARMONIE 10 5 MMIJ LIDAR 10 0 9 54 Wind speed m s nra a een 10 100 Heiaht Iml Wind speed versus height OWEZ HARMONIE Height m Figure 1 Tall mast validation results in red KNW data in blue cup and in green sonic anemometer measurements but for MMIJ green means LIDAR FINO1 100 m MMIJ 90 m hourly wind speed m s Beaufort scale Beaufort scale hou
4. or in minutes and seconds 54 27 00 N 2 58 12 E In the RD coordinate table the value is 2041 720640 which means x 2041 meters and y 720640 meters If you open the table in a spreadsheet such as EXCEL the spreadsheet might interpret the as a division and display the result but the coordinates can still be seen if you click on the spreadsheet cell of interest https nl wikipedia org wiki Rijksdriehoeksco C3 B6rdinaten 3 2 Image browser On the website there is an image browser that can be used to view images of the dedicated KNW datasets These dedicated datasets are based on the same information as the basic KNW datasets but have either been used to calculate values that are useful for wind energy purposes and or are extracted for a specific location More information concerning the datasets can be found in section 4 1 about the downloadable END USER PACK which contains these datasets both the imagery and underlying data The images that can be viewed are e 2D figures for each of the 8 standard heights of the KNW atlas and covering the whole of the KNW domain average wind speed Weibull parameters maximum wind speed min max 10 90 of 11 year running averages and the once in 10 50 and 100 year extremes e figures wind speed distributions wind roses and time series and TAB files for the 12 selected location grid points listed here below The image browser can be
5. 013 and 2004 2013 and for other heights More precisely all the 1D images and TAB files listed in section 3 2 can be found in this folder Figure 10 is an example of a TAB file in this case for the KNW grid point horizontally closest to Meetmast IJmuiden MMIJ at a height of 100 m the values at the grid points directly above and below this height are linearly interpolated and for the whole 35 year period It shows the wind speed frequency distribution for 12 wind direction sectors and for all wind directions together It also shows the frequency distribution of the wind direction sectors for 30 wind speed bins of 1 m s The last column shows values for all wind directions combined the other columns show values for 30 degree wind direction bins The first row lists the average 14 wind speed m s the second row the Weibull Scale parameter A m s which is a related to the average wind speed but not exactly the same and the third row the Weibull Shape parameter k The value for ff gt 0 gives per wind direction sector of 30 degrees the frequency of occurrence of these wind directions considering all wind speeds The total number of data used here is 306816 N which is roughly speaking 35 years x 365 days x 24 hours In the remaining rows the TAB file shows how the data is distributed over the wind speed and wind direction bins The values in these bins are in units of 0 1 percent of all the data for the relevant wind
6. 13 14 36 38 51 53 530 48 53 63 66 53 45 350 5 4 14 15 28 23 41 45 38 35 46 59 59 43 36 37 4 5 13 16 19 15 27 34 28 28 38 55 50 37 30 28 2 7 16 17 14 11 16 19 15 16 32 47 46 32 Z1 Z2 2 9 17 18 10 6 10 13 7 11 27 41 38 27 19 16 2 3 18 19 6 gt 7 7l 4 7 21 32 26 23 14 11 it 1 8 19 20 1 3 4i 3 4 14 271 20 16 11 10 1 3 20 21 1 1 3 0 1 2 10 22 16 11 amp 3 1 0 21 22 Of 0O OF OF Of 21i 101 146 9 891 5 310 7 22 23 O D Oo D a O 5 10 5 amp 3 2 0 4 23 24 O oO OI OF OIL OF 4l 5 3 3 2 110 3 24 25 oO oO Ol OF OFF Of 31 4F 21 21 1 0 0 2 25 26 0I oO OF OF Of Of 2 2 11 2 tt of a1 26 27 Ol oO Oo OI OF Of 1i 2 oO 1i 0l 0 I 0 1 27 28 Oll 0l lIl OJ O0 0i OIJ 1 OI Ol l 0 0 0 28 29 o 0 Oo 0 O 0 1 O 0 a 0O I 0 0 29 30 0 l Oo 0 0 0 0 0 0 0 0 o 0 0 0 Figure 10 Example of a TAB file Meetmast IJmuiden at 100 m height 15 The wind speed distribution in figure 11 is based on the TAB file shown in figure 10 All wind directions have been included so just as in the last column of figure 10 we see a prevailing wind speed between 9 and 10 m s The Weibull parameters are calculated in two different ways 1 according to Wieringa Rijkoord 1983 formula 5 18 wi
7. USER PACK described in section 4 1 1 In folder GridPoints in END USER PACK described in section 4 1 hourly data for these locations can also be found in this folder Weibull parameters conform Wieringa Rijkoord 1983 formula 5 18 with a variable upper limit the bin that contains the 99th percentile of the distribution 1 Plus the Weibull parameters calculated in two different ways 1 according to Wieringa Rijkoord 1983 formula 5 18 with a variable upper limit the bin that contains the 99th percentile of the distribution and 2 conform http help emd dk knowledgebase content WindPRO2 9 03 UK WindPRO2 9 ENERGY pdf 8 KNMI North Sea Wind KNW Atlas standard domain Figure 5 The 2D geographical extent of the dedicated dataset the KNW domain left and an example of the 1D geographical extent the location grid points in the Borssele area right You can select two figures side by side to facilitate comparing one figure with another 4 KNMI Data Center The four datasets in KDC are the three basic datasets described in section 3 1 1 and the dedicated dataset relevant for most users in the END USER PACK The END USER PACK is described in section 4 1 and the imagery in the END USER PACK can also be viewed with the image browser on the KNW website which is described in section 3 2 As mentioned earlier in the manual there is a link to the KDC on the KNW website http projects knmi nl k
8. User manual of the KNMI North Sea Wind KNW atlas Version 8 September 2015 I L Wijnant A Stepek M Savenije and H W van den Brink KNMI North Sea Wind F100 i 979 2013 mean wind spee 3 4 2 6 7 54 vS 53 50 OSES 53 e9 52 50 0S 2S cS 92 51 50 OS LS 51 LS 4 0 8 0 10 0 12 0 Contents 1 KNMI North Sea Wind KNW Atlas cccccccceccccseccecesececeneceeenceeeeecsseueeeeeeueseeeneeeeeuceseueceseueeeseeness 3 2 Validation OF the KNW lt allaS eorne EEAO 3 I SS eet Sven sere cee erences E ET 5 DAISY Cae aaee E E a tuphones seeatarieuatactenuets 6 Bee Ui Se VOW S E an r beodand nrobetoe ade nseeseeeortesdaesheacie teh A 8 ARNM DU SCP coimriar ia E A A E A TA A EENE A a A TE setae sce 10 4 1 KNW KNMI North Sea Wind Atlas END USER PACK seessssssesssersssssssrseresesssssrrreeesssssseene 10 4 2 KNW KNMI North Sea Wind Atlas daily 3D model data aeeseeseesecreserrrsrrrsrrrrsrrrerrreeen 18 4 3 KNW KNMI North Sea Wind Atlas grid cell time series NetCDF eeesseserrrerrreerrrsee 18 4 4 KNW KNMI North Sea Wind Atlas grid cell time series ASCII CSV esscesceccccceecreee 19 5 How to use and not to use the KNW atlas c cccccccesecccesececeesceeenceseeseceeeeeeeeeeseeseeseueseteueeees 19 SEDUNIA CCG e E AAA 19 5 2 Wind maxima over land s sessssssnssssseressssrrrsssssererssssrrressserressssrressssrereeesssereessseereesssereesseeeeeee 21 1 KNMI North Sea Win
9. are at the top of the list and you can scroll down using the page button The selected dataset consists of 24 hourly data from 01 00 to 00 00 UTC on the day you have selected for every grid cell in the KNW domain 188 x 170 x 8 e You can download the selected day files select HTTP or FTP e Once the netCDF file is opened you will find a list of sub datasets o where in the name D is wind direction F is wind speed P is pressure Q is specific humidity and T is temperature The numbers indicate the height above mean sea level 10 20 40 60 80 100 150 or 200 m height Each of these datasets represents 24 2D maps of a weather parameter at a given height one map for each hour of the selected day o where structural information is stored e g time shows that data are stored from 01 UTC until OO UTC each day lat and lon show respectively the latitude and longitude of the 188 x 170 grid cells 4 3 KNW KNMI North Sea Wind Atlas grid cell time series NetCDF e Click on KNW KNMI North Sea Wind Atlas grid cell time series KNW netCDF TS e You can visualise the dataset before downloading click on visualize e Select the grid cell of interest from the horizontal grid of the KNW domain The grid index is included in the name of the netCDF file e g KNW 1 0_H37 ERA_NL 001 002 nc where 001 is the longitude index 001 170 and 002 is the latitude index 001 188 The lowest index values are at the top of t
10. atellite based measurements of the 10 m wind above the North Sea There is obviously a limit to this because there is a limit to how far ahead you can forecast accurately So KNW winds should not be used for diurnal analysis unless care is taken to account for this artefact An example where the KNW atlas wind speeds can still be used in an application involving diurnal analysis despite the saw tooth pattern is for the calculation of wind turbine noise Since 2004 European legislation requires distinguishing between 3 periods of the day daytime 07 19 LT evening 19 23 LT and night 23 07 LT These are then combined into one formula for the year average noise level caused by a wind turbine or park Lden Level day evening night which is Lday evening 5 L night 11 12 10 4 10 10 10 10 Liden 10 Vlog AO eet Oe The 3 L s are based on the frequency distributions of the wind speed for the corresponding 3 periods of the day 1 Source https nl wikipedia org wiki Lden In the Netherlands the difference between UTC and local time is 1 hour in the winter and 2 hours in the summer e g 18 UTC 19 LT in the winter and 20 LT in the summer 20 Windspeed F058 FO060 MMIJ HARMONIE Diurnal Cycle wipeo Sc a aT a 2 oa sai 90 gt aj 8 51 B o a a a pi tg tg oe ee L 0 2 4 6 a 10 12 14 16 18 20 22 24 Figure 16 Diurnal analyses at of the hourly average wind speed at 60 m height at MMI Measurements blue and KNW atla
11. bo mo i f SoM OMWOs Ss Figure 12 Example of a wind rose Meetmast IJmuiden at 100 m height For all selected grid points the 6 Borssele points and 6 reference points in the KNW domain described in section 3 2 time series of the yearly averages and smoothed 11 year running means are provided obviously the 11 year running mean can only be calculated for the 35 year period the 10 year period is too short as well as the average over the whole period Figure 13 shows an example for MMIJ at 100 m height for the whole 35 year period The aggregated wind speeds shown on the graph are not available in the END USER PACK but the hourly data is In the basic datasets described in paragraph 3 3 3 and 3 3 4 hourly time series are available for all of the grid points in the KNW domain F100 Average wind speed for MMIJ 1979 2013 yearly means smoothed 11 year running mean mean 1980 1985 1990 1995 2000 2005 2010 year Figure 13 Example of a wind speed time series Meetmast IJmuiden at 100 m height Yearly averages in black smoothed 11 year running mean in blue and the average over the whole period in red 17 4 2 KNW KNMI North Sea Wind Atlas daily 3D model data e Click on KNW KNMI North Sea Wind Atlas daily 3D model data e Click on visualize if you want to visualise the dataset before downloading e Select a day within the 35 year period from 1979 01 01 01 00 to 2014 01 01 00 00 The most recent days
12. ced to fit into fewer wind speed bins than for wind direction sectors with higher wind speeds MMIJ lon lat 3 44 52 85 Hws F100 Hwd D100 N 306816 15 1 15 1 45 1 75 1 105 1 135 165 195 225 255 285 1 315 ALL 15 435 75 103 135 165 195 225 255 285 315 345 I 4 2 2 5 2 3 2 3 2 2 2 1 2 1 2 4 2 4 2 2 2 1 2 2 2 2 f gt 0 6 5 6 1 6 5 6 5 53 1 4 9 7 5 15 2 114 4 110 7 8 9 7 7 00 01 10 10 8 9 13 10 a l 3 3 G 6 7 O 7 01 02 29 27 25 25 30 29 211 10 12 16 21 25 4 1 02 03 43 45 41 45 51 44 34 Z0 22 30 8 amp 8 I 39 3 5 03 04 62 I 38 55 52 67 67 47 28 33 41 49 52 4 7 04 05 79 81 64 G6 73 73 58 32 39 53 64 GS 3 8 05 06 87 84 79 751 80 3 65 41 49 64 73 76 6 7 06 07 950 90 79 B81 I 83 86 73 32 56 71 I 77 8i 7 3 07 08 93 97 83 GS 8 amp 5 85 71 I 57 62 GO 8 8Z 7 7 08 09 92 104 85 78 81 81 71 I 66 71 I 80 86 84 8 0 09 10 93 100 89 85 74 79 75 73 74 79 85 82 8 1 10 11 79 80 81 74 76 71 69 75 76 76 78 80 7 7 11 12 62 G2 75 76 75 62 66 72 74 of 70 6 amp 6 7 0 12 13 51 30 66 62 538 57 62 69 70 59 56 60 6 2
13. d KNW Atlas The KNW atlas is a 4D wind atlas based on the ERA Interim reanalyses dataset which captures 35 years 1979 2013 of meteorological measurements and generates every 6 hours 3D fields on a horizontal grid of 80 km which are consistent with these measurements and the laws of physics This dataset is downscaled using the State of the art weather forecasting model HARMONIE which generates hourly data on a horizontal grid of 2 5 km The wind speeds were then tuned to match the measurements made at KNMI s 200 m tall meteorological mast at Cabauw by increasing the vertical shear of the horizontal wind speed by 15 The same wind shear correction factor was applied uniformly throughout the whole KNW atlas domain and period The result is a high resolution dataset of 35 years the KNW atlas 2 Validation of the KNW atlas The KNW atlas has been validated against publicly available wind measurements from three tall offshore wind masts OWEZ FINO1 and MMIJ Meteorological Mast IJmuiden The difference between the measured wind speeds averaged over periods when the mast measurements were undisturbed by nearby wind parks and the KNW values is less than 0 2 m s for all masts and all measurement heights figure 1 The validation results imply that the accuracy of the long term average wind speeds of the KNW atlas is comparable to that of the measurements cup and sonic anemometer and LIDAR The same can be said of the estimates of the once in
14. d turbine noise level calculations for all periods of the day and in all seasons without loss of accuracy except for the winter evenings and nights The accuracy of the KNW atlas wind speeds should be verified for these periods before using them for wind turbine noise level calculations 5 2 Wind maxima over land The high values for the maximum hourly average wind speed at 100 m height over Groningen and the Dollard area in figure 6 repeated in figure 17 top are from an extreme event that the KNW atlas read HARMONIE produced for the 1 of August 1983 06 UTC with a chance of happening of once in 10 000 years There are few if any measurements at 100 m from this period to compare the KNW data to so we use measurements at the standard meteorological wind measurement height of 10 m for the comparison The measured wind speeds are unusually strong but cannot be classified as extreme because they occur more than once a year The KNW maximum hourly average wind speeds at 10 m height figure 17 below are extreme and significantly 21 higher than the ones measured that day Where the KNW atlas produces a daily maximum of the hourly average wind speeds of more than 18 m s between 05 and 06 UTC the nearest KNMI measurement site Eelde 280 measures a daily maximum of only 11 8 m s between 11 and 12 UTC which is much more realistic time of day for the maximum wind speed in convective weather situations than between 05 and 06 UTC Deelen 275 a
15. es are available for 5 weather parameters temperature wind speed and direction pressure and specific humidity and the units are included in the file Each daily file is for a given height above mean sea level 10 20 40 60 80 100 150 and 200m are available Considering the 3D files for each of the 8 heights together provides a 4D description of the lower atmosphere The complete set consists of 35 years of 365 366 days 2 KNW KNMI North Sea Wind Atlas grid cell time series KNW NetCDF TS here you will find NetCDF files organised per location 188x170 grid cells with central point in lat lon in the KNW domain you can select a grid cell location and download all available time series 35 years for that location in one go Output variables are temperature wind speed and direction pressure and specific humidity on 8 levels 10 20 40 60 80 100 150 and 200m 3 KNW KNMI North Sea Wind Atlas grid cell time series KNW CSV TS the same as in previous set but then in ASCII CSV files The basic daily dataset is fairly large 188x170 domain size x 24x365x35 hourly output over 35 years x 8 levels x 5 variables which is in total about 400 x 10 numbers 1 5 terabytes The dataset with ASCII CSV files is smaller 0 75 terabytes A more detailed description of these datasets can be found in sections 4 2 4 4 There is a where filter function in KDC which is particularly useful if you want to find a t
16. found on the KNW page of the KNMI website http projects knmi nI knw under Image Browser blue top bar The selection scroll bars at the top of the page can be used to select 1 The geographical extent of the required information either the KNW domain 2D or a location 1D The 12 locations are e 6 grid point locations in the KNW domain figure 5 left closest to Cabauw Europlatform 1 FINO 1 MMIJ Meetmast IJmuiden OWEZ Windpark Egmond aan Zee and Lichteiland Goeree 6 grid points in the Borssele area figure 5 right Borssele point O BORSO 1 BORS1 2 BORS2 3 BORS3 4 BORS4 and 5 BORSS 2 The information you require note that the number of dimensions have to match those of the geographical extent already selected 2D for the area of the KNW domain and 1D for a location e Average wind speed 2D e Weibull shape and scale parameter 2D e Maximum wind speed 2D e Min max 10 90 of 11 year running averages 2D e Once in 10 50 and 100 year extremes 2D e TAB files 1D e Wind speed distribution 1D e Wind rose 1D e Time series of yearly average smoothed 11 year running average only available when the period selected is the 35 year period and period average wind speeds 1D 3 The level 10 20 40 60 80 100 150 or 200 m height 4 The period 35 years 1979 2013 or a subset of 10 years 2004 2013 Figures and underlying aggregated data in folder 2D domain in END
17. he list and you can scroll down using the page button above the list of files The selected dataset consists of hourly data see the 5 weather parameters and 8 heights listed in the next bullet point for the period 1979 2013 e You can download the selected grid cell files select HTTP or FTP e Once the netCDF file is opened you will find a list of sub datasets o where in the name D is wind direction F is wind speed P is pressure Q is humidity and T is temperature The numbers indicate the height above mean sea level 10 20 40 60 80 100 150 or 200 m height Each of these datasets represents a single time series of a given weather parameter at a given height o where structural information is stored e g time shows that data are stored for 306816 hours from 19790101 at 01 UTC up to and including 20140101 at 00 UTC lat and lon show respectively the latitude and longitude of the grid cell For software that you can use to manipulate or display NetCDF files we refer to www unidata ucar edu 16 For software that you can use to manipulate or display NetCDF files we refer to www unidata ucar edu 18 wind speed i LE Lal B METET A li Ny Hh fi N hy Mh HI Nh wind speed m s 1 o Hl 0 0 i gt p ypy ypy o y yy y y y y l y y o aas 1979 01 01 1982 11 21 1986 10 11 1990 09 01 1994 07 22 1998 06 12 2002 05 02 2006 03 23 2010
18. he basic datasets to facilitate script controlled downloading of the large number of files The supplementary files can also be found in the END USER PACK in the KDC http projects knmi nl knw data The links under Grid point locations on the website are particularly important They contain the conversion tables that you will need to translate the coordinates of the location of interest in lat lon or RD coordinates into the KNW atlas index of the corresponding grid cell The index of the grid cell is the last part of the filename containing the downloadable time series for that grid cell In this example KNW 1 0_H37 ERA_NL 040 180 zip the 040 is the part of the grid cell index with values from 001 to 170 that shows how far east the grid cell lies and the 180 is the part of the index 001 188 that shows how far north so this grid cell lies in the far north of the KNW domain The conversion tables have 188 rows and 170 columns The east west index corresponding to longitude and the RD coordinate x increases from left to right up to 170 as one would instinctively expect The north south index is the opposite to what one would expect it increases from top to bottom up to 188 with the low values at the top corresponding to the north end of the KNW domain and the high values at the bottom corresponding to the south The value in the lat long conversion table in row 180 and column 040 is 2 97 54 45 which means 54 45 N 2 97 E
19. hows examples of the average and maximum wind speed at 100 m height for the whole period of the KNW Atlas Maps of the period minimum wind speed are also available Note that the wind speed legend is different for the average and the maximum wind speed The maximum wind speed is the highest value of the hourly average wind speed in the 35 years that the KNW dataset covers The maximum values over sea are an accurate representation of the wind speed on the hour and day when the maximum occurred in reality The high values over Groningen and the Dollard area are less accurate They are from an extreme event that the KNW atlas read HARMONIE produced for the 1 of August 1983 This value is theoretically possible once in 10 000 year event but the wind speeds measured on that day were significantly lower see chapter 5 2 11 KNMI Nagin Sea Wind A100 1979 2013 eibul scale ee 3 4 5 6 KNMI N a Wind k100 1979 2013 iai shape eter 3 4 5 6 Figure 7 Example of the scale left and shape right parameter right of the Weibull distribution at 100 m height for the whole KNW domain and the whole 35 year period Weibull parameters examples in figure 7 have been calculated conform Wieringa Rijkoord 1983 formula 5 18 with a variable upper limit 0 5 m s bins from 4 m s up to but not including the bin where the cumulative distribution equals or exceeds 0 99 12 aia maxim ENMI iirmmini 1 201 MIMI N 1 1 1 201 NOR ear unni Pip rm nmi
20. imeseries for a specific location It enables you to zoom in on the area of interest Choose coordinates and four boxes will pop up In these boxes you can fill in the lat lon coordinates that define the boundaries of your area to the west longitude left box the east longitude right box the south latitude lower box and the north latitude upper box E g for location Spijkenisse 4 31 left 4 33 right 51 82 lower and 51 84 upper Then click on FILTER In KNW CVS TS 3rd row you will now find 1 page instead of more than 6000 pages 6 Latitude x longitude x time x height There are 188x170 grid cells in the KNW domain For grid cell to lat lon conversion see http www knmi nl samenw knw data GridPointLocations_lonlat without applying the where filter with timeseries for the 3 gridpoints within the area selected One of them corresponds best with the location of Spijkenisse 079 065 51 84 N 4 33 E 3 1 2 Supplementary Files On the KNW webpage http projects knmi n l knw data you will also find supplementary files with information relating to domains location grid points and file listings of the basic datasets to facilitate script controlled downloading e Maps of the KNW and Borssele domains and the position of the 12 location grid points e Latitude longitude and RD coordinates of all of the KNW grid points in ASCII CSV Excel or NetCDF e File listings of t
21. n K N yea om rmmax 00 1979 2013 a 6 3 i 4 0 6 0 8 0 10 0 12 0 4 0 6 0 8 0 10 0 12 0 yi ion 1 2 KNMI 1 year fanning hea rm ala KNMI iii yin in 20h pore 2013 10 0 12 0 Figure 8 Example of the minimum upper left maximum upper right 10 percentile bottom left and 90 percentile bottom right of the 11 year running average wind speed at 100 m height for the whole KNW domain and the 35 year period Calculated with climate data operators 13 KNMI North Sea Wind U10y100 1979 2013 KNMI North Sea Wind U100y100 8a e018 10 year return Spee 100 year return Spee 3 4 5 6 r 3 5 6 7 52 52 50 53 53 50 54 51 50 Figure 9 Example of the once in 10 year left and once in 100 year right extreme wind speed for the whole KNW domain based on a Gumbel fit of the annual maxima of 1979 2013 Once in 50 year extreme is not shown The parameters of the Gumbel distribution are determined by Maximum Likelihood from the annual maxima The return value RV for a given return period T is given by RV u o0o ln T Here wis the location parameter and o the scale parameter of the Gumbel distribution 4 1 2 Grid points folder Figures 10 13 illustrate the images and aggregated data available in the Grid points folder for the 12 selected grid points using location Meetmast IJmuiden MMIJ and the period 1979 2013 as an example The images and aggregates are available for both 1979 2
22. nd Beek 380 register higher values for the daily maximum than Eelde respectively 12 3 m s and 13 9 m s both between 08 and 09 UTC but still significantly lower than the KNW maxima We want to emphasise here that the maximum values in the KNW atlas above land are possible the 100 m maxima near Groningen occur once in 10 000 years but may be much higher than what has actually been measured in the past 35 years The maximum values over sea areas are much more accurate KNMI North Sea Wind ffmax100 i 979 2013 maximum wind spee 0 0 12 5 25 0 37 5 50 0 Figure 17 Example of the maximum wind speed at 100 m height top and at 10 m height below for the whole KNW domain and the whole period of 35 years 22
23. nie F100 climate 2004 2013 EENE 7 50 825 Grid cell 1 1 Grid cell 1 0 1 1 s 1 51 5043 7 45 50 29 SZ ES See KNW webpage GridPointLocations lonlat csv excel KNW data supplementary files es gt E mnia wne a a 5250 53 25 al EEE 05 26 Sits uy pe vo O amp e a 7 a oo mc 51 75 51 IS a a aad AEEA HAE N Grid cell 1 188 eS MRR ric coll 170 188 re a a ae a a a a L46 54 63 170 grid points 01 448 50 25 Figure 4 Left panel KNW data have been calculated within the blue frame and are freely available in the coloured area the KNW domain The grid lines indicate the ERA Interim grid cells 80 by 80 km Middle panel zoomed in on the KNW domain Right panel grid cell point locations within the KNW domain the layout is the same as the conversion file on the KNW webpage GridPointLocations_latlon csv but is upside down compared to the maps in the other panels Part of the data is stored in the KNMI Data Centre KDC https data knmi nl portal KNMI DataCentre html but the link to KDC can be found on the KNW webpage It is however possible to make these data available but only under certain financial conditions for more information https secure knmi nl contact contact php gt The grid cells 188 are the ones with the highest latitudes furthest north and the grid cells with 1 the ones with the l
24. nw data Use filter term KNW to get to the datasets described in the rest of this chapter 4 1 KNW KNMI North Sea Wind Atlas END USER PACK e Click on KNW KNMI North Sea Wind Atlas END USER PACK e You can visualise the dataset before downloading click on visualize e You can download the whole END USER PACK in one go select HTTP or FTP The END USER PACK contains zip files grouped in 3 folders 2D domain GridPoints and Supplementary The zip files have to be extracted unpacked before they can be viewed e g with Paint or Irfanview or used for further analysis 4 1 1 2D domain folder Figures 6 9 show examples of the imagery contained in the dedicated datasets and the 2D domain folder of the END USER PACK All of the examples are for 100 m height and 1979 2013 but maps for 2004 2013 and other heights are also available More precisely all the 2D images listed in section 3 2 can be found in this folder The aggregated data used to make the images can also be found in the 2D domain folder 1 j KNMI Data Center 10 KNMI North Sea Wind F100 on 979 2013 KNMI North Sea Wind Le bac 1979 2013 mean wind spe maximum wi 3 4 5 6 7 3 4 5 6 m s nk Ha 4 0 6 0 8 0 10 0 12 0 0 0 12 5 25 0 37 5 50 0 Figure 6 Example of the average left and maximum right wind speed at 100 m height for the whole KNW domain and the whole 35 year period Figure 6 s
25. owest latitude furthest south The grid cells 1 are the ones with the lowest longitude furthest west and 170 with the highest longitude furthest east 3 1 KNW data 3 1 1 Basic KNW datasets KNW data are stored in the KNMI Data Centre KDC and there is a link on the KNW website that gives you access to KDC You will find the link to KDC if you select KNW data blue top bar Once in the KDC you can use filter term KNW the which filter function to get to the model data which has been organised into the three basic datasets The datasets can be downloaded but before doing so please read this manual and make use of the conversion tables see links under Supplementary files Grid point locations on the website and described in 3 1 2 You need the conversion tables to translate the coordinates of the location of interest in lat lon or RD coordinates into the KNW atlas index of the corresponding grid cell The three basic datasets contain exactly the same data but are organised differently or in different formats to suit different users 1 KNW KNMI North Sea Wind Atlas daily 3D model data KNW NetCDF 3D here you will find NetCDF files organised in daily files for a fixed domain the KNW domain described in figure 4 You can select a day and download all available information for that day 23 values from 01 until 23 UTC and one from the next day the 00 UTC in one go The daily fil
26. rections 10 1 2010 easterly 13 1 2010 southeasterly and 16 1 2010 southerly FINO 100m hourly 25 observations HARMONIE wind speed on 10 01 11 01 12 01 13 01 14 01 15 01 16 01 17 01 18 01 19 01 20 01 date Figure 15 Comparison of measured hourly average wind speeds blue and KNW values red for FINO1 at 100 m height from 10 until 20 January 2010 Wind speeds in m s However there is a problem present in the time series of figure 15 but it is so small that it is barely visible Because of the way the KNW atlas was constructed the atlas reconstructs the diurnal variation of wind speeds less accurately than wind speeds averaged over longer periods The saw tooth patterns exhibited by the graphs of the diurnal variation example in figure 16 are an artefact of the KNW atlas This is caused by the initialisation of the HARMONIE weather forecasting model at 0 06 12 and 18 UTC with the lower ERA Interim wind speeds It takes a number of hours into the forecast before HARMONIE develops more realistic smaller structures and higher wind speeds than ERA Interim is capable of The forecast hours included in the wind atlas dataset hours 1 6 with O being the initialisation time were chosen to optimally match measurements made at KNMI measurement sites at the meteorological standard height of 10 m Since then however Wijnant et al 2015 have shown that using longer forecast lead times would probably improve agreement with s
27. rly wind speed m s observations e HARMONIE e 0 1 0 2 0 5 1 2 3 4 5 10 0 1 0 2 0 5 1 2 3 4 5 10 return period year return period year observations HARMONIE e Figure 2 Tall mast validation of extremes based on measurements red and KNW data blue for FINO1 100 m left and MMIJ 90m right The hourly wind speeds in the KNW dataset represent 40 60 min averages page 6 TR352 More information e Tall mast validation http www knmi nl bibliotheek knmipubTR TR352 pdf e Scatterometer validation http www knmi nl bibliotheek knmipubTR TR353 pdf Harmonie minus qscat m s 5 0 9 0 7 0 5 0 3 0 1 0 1 0 3 0 5 0 7 0 9 5 son gt 54 N Figure 3 Scatterometer validation showing average 10 m wind speed difference bias between KNW atlas and QuikSCAT 25 km product 3 KNW website The wind climatology that is freely available is the KNW data in the coloured area in figure 4 referred to as the KNW domain The domain s latitude roughly from north to south consists of 188 grid cells and the domain s longitude roughly from east to west of 170 grid cells All freely available data can be found on the KNW website http projects knmi n knw KNW data have also been calculated outside this area within the blue frame of figure 4 but these data are not freely available Model domain cutout of Harmonie simulation with ERA Interin gr ERA Interinv Harmo
28. s red KNW atlas contains the 1 to 6 HARMONIE forecasts shear correction initialised every 6 hours 00 06 12 and 18 UTC using ERA Interim values To see if the KNW atlas wind speeds can be used for this purpose whilst retaining the levels of accuracy shown for analyses of longer periods one has to compare the periods used to the periods of the saw tooth pattern e The daytime period in winter covers exactly 2 successive saw tooth periods so the accuracy of the frequency distribution should be very similar to the accuracy of the KNW atlas as verified in earlier reports The change due to summertime should be small e The evening period is slightly shorter than 1 saw tooth period and the missing forecast hour in summer is the fourth which is close to the measured wind speed because it is one of the middle hours in the saw tooth period In winter the fifth hour is missing so the frequency distribution will have a larger bias to lower wind speeds than when the KNW wind speeds are analysed for longer periods e The night period is slightly longer than the saw tooth period and the consequences are unimportant because the extra forecast hours are the third fourth and fifth However in the winter the extra forecast hours are the fourth fifth and sixth so the frequency distribution will have a larger bias to higher wind speeds than when the KNW wind speeds are analysed for longer periods So the KNW atlas can be used for diurnally dependent win
29. th a variable upper limit the bin that contains the 99 percentile of the distribution and 2 conform http help emd dk knowledgebase content WindPRO2 9 03 UK WindPRO2 9 ENERGY pdf The 1 m s bins of figure 11 were chosen to make the graph clearer but smaller bins of 0 5 m s were used when calculating the Weibull parameters F100 Wind speed distribution for MMI 1978 2013 100 77 Weibull Wieringa parameters A 11 1 k 22 Weibull WAsF parameters A 11 0 k 2 1 rite speed m a Figure 11 Example of a wind speed distribution Meetmast IJmuiden at 100 m height Again we show the KNW climatology for grid cell MMIJ at 100 m height that was presented in figures 10 and 11 but this time in the form of a wind rose figure 12 The wind direction bins are the same as in the TAB file figure 10 but the wind speed bins are chosen differently 3 m s instead of 1 m s bins to make the graph clearer The percentage of the data shown by the various colours is also different in this case they are percentages relative to the total number of data and not just the data with wind directions in the given wind direction sector These aggregates have not been included in the TAB file It would be difficult to misinterpret this figure as opposed to figure 10 and conclude that the most prevalent combination of wind speed and wind direction was one of the NNE bins 16 F100 WindRose for MMIJ 1979 2013 LETS g E Po
Download Pdf Manuals
Related Search