RESULTS OF FOUR YEARS MONITORING OF DISPLACEMENTS OF ROCK BLOCKS IN THE KRKONOŠE MTS.

Transcription

1 Acta Geodyn. Geomater.Vol.1, No.3 (135), 75-81, 4 RESULS OF FOUR YEARS MONIORING OF DISPLACEMENS OF ROCK BLOCKS IN HE KRKONOŠE MS. Otakar ŠVÁBENSKÝ* and Josef WEIGEL Brno Unversty of echnology, Department of Geodesy, Veveří 95, Brno, Czech Republc el.: , , Fax.: *Correspondng author s e-mal: svabensky.o@fce.vutbr.cz (Receved March 4, accepted May 4) ABSRAC he epoch measurements of dplacements of the rock blocks started n n western part of Krkonoše Mts. he local area of nvestgatons n vcnty of Labská meadow and Volík was covered by three sngle geodetc networks named A, B, C. he measurng actvtes were ntated by KRNAP (the Krkonoše Mts. Natonal Park Admnstraton) n connecton wth the research project concernng the complex analyss of long-term changes of the Krkonoše mountan tundra wth specal am to nvestgate the relevantly mportant hydro-geologcal structural elements. he paper presents detaled descrpton of the sngle networks and the employed measurng technologes. Besdes GPS also other classcal geodetc measurng technques (precse levellng, angle and dstance measurements) were used for montorng of the relatve horzontal and vertcal postonal changes. GPS campagns were repeated after year, epoch nterval for heght measurements after half a year. In 3 the network C was extended by addtonal ponts at rock blocks stuated on southern slopes of Volík, where greatest possble changes should be expected. he paper also presents some results of geodetc measurements coverng four years of dsplacements montorng whch ndcate possble movements of some measured blocks mostly n vertcal drecton. KEYWORDS 1. geodesy, deformaton montorng, geodynamcs INRODUCION In upper parts of Krkonoše Mts. a phenomena called ploughng stones s often detected (Kocánová, 1999). Man cause for the slow moton of solated rock blocks s obvously the frost nfluence. In slope drecton a small mound ahead of each block s formed, and behnd s a depresson (hollow). Such blocks are spread over the Labská - Pančavská Meadows area n great number. Dmenson of the Fg. 1 Ploughng Block Scheme and Pcture blocks dffer from.5 m to about.5 m, and most of them are covered by vegetaton grass, moss, eventually dwarf pne. A typcal block s shown n Fg. 1. It s to be expected that the extreme atmospherc condtons prevalng n the mountans have essental nfluence also on the movements of boulders n rock-block felds. here exsts an opnon that the moton velocty s dfferent for free rockblock felds and for those covered by vegetaton.

2 76 O. Švábenský and J. Wegel able 1 Approxmate Coordnates of Prncpal Ponts Pont Lattude Longtude Heght ZLNA 5 o 45 4, 15 o 3 1, 1395, m A1 5 o 46 1, 15 o 3 1,1 1359,9 m B1 5 o 46 6,8 15 o 3 1,7 148, m C1 5 o 46 39,5 15 o 3 4,8 146, m Specalsts from the Krkonoše Mts. Natonal Park Admnstraton are also studyng these phenomena n scopes of varous grant projects. In cooperaton of KRNAP and Brno Unversty of echnology, Department of Geodesy, three areas n western part of Krkonoše Mts., n localtes of Labská Meadow and Volík were chosen. In each of the three areas a certan number of stones had been selected for the purpose of long-term montorng of the eventual dsplacements. Denomnatons of the areas are A, B, and C, wth elevatons from 135 m to 143 m above sea level. Average number of blocks n each area s about 1, each stone has specal precse small screw marker for precse centrng of geodetc nstruments. In ths way the three ndvdual local geodetc mcronetworks were establshed, and n the perod 3 they were subsequently montored n 7 measurng campagns. Connecton of the three networks to the superor geodetc reference systems was accomplshed by GPS method.. GEODEIC NEWORKS AND MEASURING MEHODS hree geodetc montorng networks were establshed n the terrtory of nterest, n areas named A, B, and C. Area A s located at Labská Meadow range, cca 3 m to the north of crossroad U čtyř pánů, on the rght sde of the tourst path leadng to the sprng of rver Labe. he range s covered by subalpne meadow wth sparse growth of dwarf pne, on the 6% slope n south-east drecton. In eght selected rock blocks were marked wth specal geodetc screw markers (A1 - A8). he stones are not standng out very much above surroundng terran, and most of them are partly or completely overgrown wth grass. Excepton s the bg block (A8) whch s about 1. m hgh. Area B s located southwards to the Labe sprng, on both sdes of the tourst path. erran coverage s the same as n the area A, on the 8% slope n southeast drecton. Seven rock-blocks, all of them covered wth grass, were marked here wth screws (B1 - B7). Area C s located northwards to both prevous areas, n close vcnty of state fronter lne between Czech Republc and Poland. It s large rock-block feld located 5 m southwards to the rocky top of Volík. he terran slope s from 1% to 3% here, n south drecton. In frst four blocks were marked here, and now (3) there are ten marked blocks (C1 - C1). Eght of them are spread on a horzontal lne whch goes through the whole vast rock-block feld where greatest possble changes should be expected (Demek 1991). he markers are alternatvely on bare and overgrown blocks. In each network a sngle man pont had been selected, whch was then connected to the reference pont ZLNA by statc GPS method. he pont ZLNA s ted to state coordnate system ERF-89 over the pont Žalý, whch s ncluded n the natonal satellte network DOPNUL. Approxmate coordnates of these mportant ponts are gven n able 1. he heghts are gven n the state vertcal datum Bpv. he vertcal dsplacements were montored by very precse levellng method. In each of the networks ndvdual sngle heght dfferences betwen ponts were measured so that the loops could be formed and checked by the loop closures. An average loop closure was.45 mm, an accuracy of a sngle heght dfference was about.1. mm. he horzontal dsplacements were determned by trgonometrc method, wth measured angles and dstances. Drectons were measured by precse optcal nstruments Zess heo 1A, and dstances were measured by parallactc method wth m subtense bar. An accuracy level of horzontal postons of ponts s about 1 3 mm. Sx measurng campagns were carred out here n the past four years. he lst of these campagns together wth an overvew of measurng technques employed s gven n able. As can be seen, the most frequent technque s precse levellng, whch was used n every measurng epoch. In October the angle and dstance measurement could not be carred out because of very unfavourable weather condtons, whch was also the case n September PROCESSING OF HE ERRESRIAL DAA Fnal coordnates and heghts of the network ponts n each measurng epoch were determned by parametrc adjustment based on least squares method. Computatons were carred out wth the help of software G-NE whch permts to work wth horzontal and/or vertcal networks. he software s very flexble when workng wth nput data stemmng from varous dfferent methods of recordng (use of dfferent measurng nstruments for data collecton). Horzontal and vertcal adjustment s accomplshed n separate steps, whch leads to less demandng computatons and to easer testng and checkng for

3 RESULS OF FOUR YEARS MONIORING OF DISPLACEMENS 77 able Hstory and Structure of Campagns Campagn 9/ 1 5/1 1/1 3 1/1 4 1/ 5 1/ /3 Method Network Horzontal Levellng GPS A Yes Yes Yes B Yes Yes Yes C No No Yes A No Yes No B No Yes No C No No No A Yes Yes Yes B Yes Yes Yes C No No Yes A No Yes No B No Yes No C No No No A No Yes Yes B No Yes Yes C No No Yes A No Yes No B No Yes No C No Yes No A No Yes Yes B No Yes Yes C Yes Yes Yes outlers. Output contans not only the lst of adjusted coordnates and heghts, but also the detaled protocole documentng the acheved accuracy and other mportant attrbutes of nput data and adjustment process. Here belong.e. mean square errors of coordnates/heghts and parameters of the error ellpses for determned ponts. he software enables to perform a constraned or free adjustment, and t s possble to modfy the adjustment procedure wth dfferent levels of constranng. Most frequent measurng method used was the very precse levellng. Because there were no possbltes to establsh fxed benchmarks n the areas nor n ther vcnty, a specal processng method for vertcal stablty evaluaton of the network ponts had been used. (Štroner 1999), (hon 3). Frst step s the selecton of dentcal levellng loops (LL), whch were measured n the same confguraton, at least n two measurng epochs. he second step s the adjustment of the loops by parametrc method wth addtonal condtons. he parameters are the sngle measured heght dfferences, and the unknowns are the heghts of all loop ponts. o remove the sngularty s necessary to ntroduce here the condton n form of n =1 H = const. (1) where H s the heght of -th pont. In matrx form B dh + b =, where B s the lnearzed matrx of condtons, and b s the vector of closures n condton equatons. Fulfllng the condton (1), the matrxes have form B = ( ), b =. () Soluton of the parametrc adjustment wth addtonal condtons s A PA B dh = A PA k B B dh A Pl = k + b, (3) B 1 A Pl, (4) b where k s the vector of arbtrary unknowns. Resduals of the sngle heght dfferences are then computed after the formula v = AdH + l. (5) Inverted matrx N 1 A PA = B B 1 Q = Q HH kh Q Q kh kk (6) s the same as the resultng matrx of weght coeffcents. hrd step s the evaluaton of stablty of sngle loop ponts between two epochs wth help of the dfference lmt. It s possble to calculate the dfference value d (left upper ndex denomnates the epoch) for any pont heght by d = ( H + H H H ) 1 n 1 d n H = H H, (7) and analogcally to calculate. Lmt dfference of value = d 1 d for assessment of stablty of sngle ponts s = t σ od + σ1d, where t s confdence factor, and σ od, σ1d are the root mean square errors of the dfferences. If no pont s outsde the calculated dfference lmt, then there s no ground to assume any movements n the loop. In the other case, when some or more of the ponts are outsde the calculated dfference lmt, the most outlyng pont s excluded, and then the computaton s repeated, wth approprate modfcaton of the formulas. Computatons are stopped when all the remanng ponts are wthn the gven dfference lmt. he example of vertcal stablty evaluaton of ponts from sx measurng epochs n area A s gven n able 3.

4 78 O. Švábenský and J. Wegel able 3 Vertcal Stablty Evaluaton of Ponts n Area A Levellng Between Ponts Loop Epochs A1 A A3 A4 A5 A6 A7 A8 1,,8,3 3 yes yes yes yes 1,3,4,5 3 yes - yes yes yes yes yes yes - yes 1,6,,8 1 5 yes yes yes - no 5 yes yes yes - yes 1 4 yes - yes yes yes - - yes 1,5,4,3,8 1 5 no - yes yes yes - - yes 4 5 yes - yes yes yes - - yes,8,3 1 - yes yes yes # of Assessments 8x 5x 6x 4x 4x 3x x 8x Postve 7x 5x 4x 4x 4x 3x x 7x Negatve 1x x x x x x x 1x yes - stable pont < použty jako vztažné yes* - fulflled < after excluson of most outlyng pont no - non-stable pont > possble movement Vertcal adjustment of the pont heghts was carred out wth the software G NE. o quanttes whch are enterng the adjustment process (.e. heght dfferences) the a-pror mean square errors have to be assgned. hese were derved n relaton to the number of statons n. σ = σ n =,mm n, h s where σ h s standard devaton of the total heght dfference, and σ s s standard devaton of the sngle heght dfference. Levellng networks at areas A, B were adjusted as free n all the measurng epochs, wth respect of the prevous stablty evaluaton of ponts. As far as the area C had been completely measured only n the last epoch (August-September 3), there was no possblty to perform any stablty evaluaton here. In Fg. the vertcal dsplacements of ponts A1, and A4 from seven levellng epochs are shown. he pont A1 dsplays possble slow snkng, whle the pont A4 behaves dfferently, wth ndcatons of some perodcal changes. A1 - Vertcal changes A4 - Vertcal changes Dsplacement [mm] 3,,5, 1,5 1,,5, -,5-1, -1,5 -, -,5-3, Epoch [year] Dsplacement [mm] 3,,5, 1,5 1,,5, -,5-1, -1,5 -, -,5-3, Epoch [year] Fg. Vertcal Dsplacements of Ponts A1, A4 (Levellng)

5 RESULS OF FOUR YEARS MONIORING OF DISPLACEMENS PROCESSING OF HE GPS DAA GPS measurements were carred out by statc method wth long observaton tmes up to 4 hours sessons. In Krkonoše campagns the Leca GPS recevers SR99/399, and SR5 were only used. Data were frst checked and then processed accordng to unversal scheme for all campagns. Processng parameters were 1 degrees elevaton mask, computed onosphere model from reference staton data, and standard Saastamonen troposphere model no correctons were estmated because of short baselnes (under 3 km) and small relatve heght dfferences between the reference and the ponts n montored areas (under 5 m). Oblgatory was the use of precse ephemerdes from CODE (Unversty of Berne). he data were prelmnarly processed wth the help of standard software LEICA SKI-Pro ver. 3., and the fnal soluton was obtaned wth the Bernese GPS Software ver. 4.. In course of prelmnary processng analyses there were some ndcatons that possble dcrepances n antenna phase centers defntons exst. he recever/antenna SR 99/399 has an nternal Ball element, lght external A5 antenna wth Aero element s desgned for use wth recevers LEICA System 5. he approprate phase center offsets (PCO), and elevaton dependent correctons patterns (PCV) are ncluded n standard Leca SKI-Pro processng software. PCO and PCV values are obvously determned by nternal Leca testng procedures. Alternatve processng wth the use of PCO/PCV sets determned by NGS (NOAA) also gave not qute satsfactory results. It was the reason for new own mutual calbraton of the antennas used. GPS antennas test base on the roof of BU faculty buldng was completed n. It conssts from 5 concrete pllars equpped wth forced centrng heads, and the staton UBO wth fxed choke rng antenna (RM 9659.) mountng. Relatve postons of the base ponts were determned by precse angle and dstance measurements. Adjusted postons of ponts were transformed nto local horzontal system. he relatve horzontal accuracy σ x,y s better than,3 mm, and vertcal accuracy σ h s better than, mm. In able 4 the results of the new calbraton of the A5 n respect to SR99/399 stemmng from extensve testng (65 days, over 35 hours of data) are gven. able 5 shows the results of 4 hours testng of relatve vertcal PCO components for combnatons of some of the Leca and rmble antennas. able 4 Leca PCO Values of A5 n Respect to SR99/399 (n m) PCO orgnal values dfferences new values [m] L1 L L1 st.dev. L st.dev L1 L dn de du able 5 estng of Relatve Vertcal PCO s of Leca and rmble Antennas (n m) Antenna ype Freq Heght PCO (NGS) Heght PCO Dfference Heght PCO (New Values) Leca SR99/399 L1.184 reference.184 Internal L.1 reference.1 rmble.+gp L Compact L1/L L rmble.-gp L Compact L1/L L rmble L Zephyr L

6 8 O. Švábenský and J. Wegel LEICA INERNAL SR99/SR399 nt. antenna ( 4) 96/6/ Fg. 3 Antenna Fle Format (IGS_1.pcv) SAR OF ANENNA RM9659. YPE / SERIAL NO FIELD Geo++ GmbH 1 7-JAN-3 MEH / BY / # / DAE. DAZI ZEN1 / ZEN / DZEN # OF FREQUENCIES IGS_3 SINEX CODE AUOMAED ABSOLUE FIELD CALIBRAION (ROBO) COMMEN G1 SAR OF FREQUENCY NORH / EAS / UP NOAZI G1 END OF FREQUENCY G SAR OF FREQUENCY NORH / EAS / UP NOAZI G END OF FREQUENCY END OF ANENNA Fg. 4 ANEX Antenna Fle Format (absolute calbraton values) he antenna PCO/PCV values are ntroduced nto the processng by way of the specal phase centre correcton fles. Fg. 3 llustrates the currently used format of an antenna fle. In Fg. 4 s an example of newly proposed ANEX antenna fle format, whch should be more convenent for correctons stemmng from the modern absolute feld antenna calbratons wth help of specally desgned robots. he values of some absolute correcton sets are avalable at ftp://gscb.jpl.nasa.gov/gscb/staton/general/pcv_pro posed/ (Schmd, et. al. 4). Besdes antenna phase center eccentrctes also other factors nfluence the GPS measurements, lke f.e multpath and dffracton of satellte sgnals. Multpath effects are caused by reflectons of sgnals from nearby surfaces, and can be dvded nto two parts. One part s nvoked by the near feld of the antenna (pllar/trpod, trbrach, adapter, marker, ground plane, etc.), second part comes from envronment. Dangerous are the frst part effects whch do not cancel out over long observaton tmes. Resduals analyss may contrbute essentally to multpath detecton. It s posslble to analyse resduals both n code and phase dfferences, takng nto account dfferent behavor of code and phase resduals. Dffracton s a phenomenon whch appears by permeaton of sgnal through lght obstructons lke f.e. folage. Both effects cause degradaton of sgnal level whch s ndcated by lowerng of SNR (Sgnal to Nose Rato) characterstcs. Such sgnals are lkely to scatter more than undsturbed sgnal. One way of dealng wth the problem s downwegtng or complete excluson of such observatons n processng procedures (Rchter, Euler 1). he approprate algorthms are ncluded even n some standard GPS processng tools, f.e. n latest versons of Leca SKI- Pro processng software. 5. CONCLUSIONS Four years of epoch montorng of postonal dsplacements of the rock block n Krkonoše Mts. dd not prove any sgnfcant horzontal changes yet. In vertcal drecton the measured dsplacements of some of the ponts ndcate possble changes wth better confdence. It s evdent that the montorng perod must be extended n tme to obtan more decsve nformaton about dsplacements trends. GPS measurements are nfluenced by many factors among whch mportant role-play the uncertantes n phase center postons/varatons of the GPS antennas, but also others lke multpath and dffracton of satellte sgnals.

7 RESULS OF FOUR YEARS MONIORING OF DISPLACEMENS 81 Probably most convenent way to the mtgaton of nconsstences n antenna phase center offsets/varatons values s the specal ndvdual feld testng procedure whch can effectvely determne the relable relatve correcton values for groups of antennae, or even for sngle combnatons of certan types of antennae. Promsng are also the results of new absolute feld calbratons wth robots, whch would be avalable n near future for many current antenna types. estng of Leca PCO values for SR 99/399 and A 5 antennas found dfference 1 mm n L relatve heght offset, whle the horzontal offsets dffer only by few mlmeters. estng of NGS (NOAA) vertcal PCO values between some of the Leca and rmble antennas revealed dfferences n L1 relatve offsets up to mm, and n L offsets up to 6 mm. Important s also the detecton and elmnaton/mtgaton of multpath effects especally for shorter observaton tmes, whch apples also for dffracton dluted sgnals. hese factors have mmnent mpact on the accuracy and relablty of resultng dsplacements, therefore t s necessary to cope wth them by means of more complete modelng, to mnmze ther proporton n the fnal dsplacement montorng results. Resduals analyss may contrbute essentally to multpath detecton. Dffracton effect s dangerous when collectng data n not optmal envronment. It can be dealt wth by use of SNR nformaton n phase data on both frequences for the downweghtng of affected observatons. Introducton of these advanced contemporary methods for mtgaton of the above mentoned dsturbng effects nto the GPS processng procedures brngs the results of dplacement montorng n Krkonoše Mts. on hgher accuracy and relablty level. ACKNOWLEDGMEN he contrbuton was elaborated wth support of the research project CEZ: J/98:6117. REFERENCES Demek, J.: 1991, Kvanttatvní výzkum pohybu sutí na svazích Hrubého Jeseníku, Geographca Geologca XXX (1991), pp Kocánová, M.: 1999, Putující kameny, Krkonošské muzeum, Vrchlabí. Rchter, B., Euler, H.-J.: 1, Study of Improved Observaton Modelng for Surveyng ype Applcatons n Multpath Envronment,. ION GPS 1 Proceedngs, Salt Lake Cty, tem5/papers/ Schmd, R., Mader, G., and Herrng,.: 4, From Relatve to Absolute Antenna Phase Center Correctons, IGS Workshop and Symposum, Berne (Draft), ects.pdf Štroner, M.:, Použtí nepřpojeného uzavřeného nvelačního posunu př měření výškových změn, Geodetcký a kartografcký obzor, č. 4/, pp Švábenský, O., Wegel, J., Machotka, R., and Podstavek, J.: 3, Recent Geodetc Actvtes of Brno Unversty of echnology n Sudety Mts. Regon, Acta Montana Ser.A- Geodynamcs, No. 4 (131), pp hon, M.: 3, Měření svahových pohybů v Krkonoších. Dplomová práce, VU Brno.