Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 1 Harmonized 2003 & 2004 XCH4 data from 15 FTIR sites applied to global validation of SCIAMACHY IMAP-DOAS v49 and WFM-DOAS v1.0: variabilities, annual cycles, biases R. Sussmann, F. Forster, T. Borsdorff, B. Dils, M. De Mazire, C. Vigouroux, T. Blumenstock, M. Buchwitz, J.P. Burrows, P. Duchatelet, C. Frankenberg, J. Hannigan, F. Hase, N. Jones, J. Klyft, E. Mahieu, J. Mellqvist, J. Notholt, K. Petersen, O. Schneising, K. Strong, J. Taylor, S. Wood taken from ESA kids web site to what detail do grown-up validation scientists believe this? Authorship and contents according to option 1: One to-be-published SCIA val paper Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 2 Zugspitze FTIR: precision of daily means is 0.4 % can retrieve true annual-cycle amplitude: 1 % can retrieve true day-to-day variability: 1 % SCIAMACHY WFMD v0.41: time-dependent biases 3 %/month due to icing problems in channel 8 could not retrieve true day-to-day variability precision of the daily means 0.6 % (calculated from the individual pixels, 1000-km selection radius) would be sufficient to capture true day-to-day variability and the annual cycle - if the icing issue could be eliminated. SCIA 2000 km daily means SCIA 1000 km daily means Zugspitze FTIR daily means * Sussmann, Stremme, Buchwitz, de Beek, Atmos. Chem. Phys, 5., 24192429, 2005 Conclusions from ACP (2005)*: SCIA WFMD v0.41 versus FTIR 2003 Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 3 non-HYMN but M. Buchwitz asked me to participate New questions/requirements for validating IMAP-DOAS v49 & WFM-DOAS v1.0 Meanwhile, the icing issue was overcome by re-implementing the SCIA retrievals from channel 8 to channel 6. Questions/requirements for validation of newest IMAP-DOAS v49 and WFM-DOAS v1.0: Question 2.1: Can the new SCIA channel 6 retrievals capture true day-to-day variability? Requirement 2.1: Make sure that g.-b. FTIR can be used as a reference, i.e., can capture the true day-to-day variability already shown by Sussmann et al. [2005]. Question 2.2: Can the new SCIA retrievals reflect annual cycles as seen by FTIR? Requirement 2.2: Check whether g.-b. FTIR really reflects true XCH4 annual cycles: investigate possible influences to the annual cycles derived from g.-b. retrievals. Question 2.3: Since the dominant source of error (time-dependent biases) has been overcome it makes now sense to perform a global validation of biases. Question: is there a latitudinal dependency of biases for IMAP-DOAS v49 and WFM-DOAS v1.0? Requirement 2.3: Make sure that CH4 columns are retrieved in a perfectly consistent manner at all globally distributed ground-based FTIR stations. Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 4 one common micro-window set (i.e., the UFTIR set) identical spectroscopic line list for all partners ? one common source of pT-input profiles ? one consistent set of a priori VMR profiles? one consistent set of regularization matrices and altitude grids? harmonized data selection / quality control? FTIR retrieval homogenization: status April 2008 Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 5 one common micro-window set (i.e., the UFTIR set) identical spectroscopic line list for all: HIT04 incl. Hase update common source of pT-input profiles: NCEP one consistent set of a priori profiles? one consistent set of regularization matrices and altitude grids? harmonized data selection / quality control? FTIR retrieval homogenization: status May 2008 Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 6 status May 2008 too low spread too high spread Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 7 FTIR retrieval homogenization: a priori profiles basis: Toon CH4 balloon profile mean tropopause altitudes derived from NCEP consistent to Remedios climatology Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 8 status May 2008 Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 9 dofs 0 dofs 0.5 dofs 1 dofs 2.5 dofs 3 dofs 4 FTIR regularization: Traditional diagonal S a - altitude-constant % VMR variabilities e.g., Rodgers and Connor, JGR, 2003 Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 10 dofs 2 dofs 2.5 dofs 3 (VMR profile scaling) dofs 4 dofs 1 FTIR regularization: Suggested Tikhonov first-derivative (L 1 ) regularization Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 11 Tikhonov L 1 profiles do by no means look worse than with diagonal S a if regularization strength is tuned to a reasonable value but there is not the disadvantage of the diagonal S a of drawing towards the a priori for low dofs FTIR retrieval homogenization: why Tikhonov L 1 : good as to profiles Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 12 AV i ( i ) ( ) over-regularized regime (i.e., dofs 1-2): slight over-estimaton of columns scatter in case of - wrong a priori profile - bad spectra (line shape) diurnal variation FTIR retrieval homogenization: why Tikhonov L 1 : more robust as to columns than diagonal S a a prioris Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 13 whatever the L 1 dofs setting is, there is per definitionem never any under-estimation of columns variability (i.e., VMR profile-scaling-type) totally over-regularized L 1 (dofs 1) is identical to VMR profile scaling. This is one of the best-tested retrieval approaches: it gives no profiles but good columns. slight over-estimation of true columns variability in case of bad spectral line shapes or bad a priori together with L 1 over-regularization reducing L 1 regularization strength (increasing dofs from 1) allows for some additional flexibility in the profile to account for i) true profile variations ii) line shape errors, iii) cloud impact on the spectra gives even more precise columns than by profile scaling FTIR retrieval homogenization: why Tikhonov L 1 : more robust as to columns than diagonal S a Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 14 FTIR retrieval homogenization: data selection / quality control strongly impacts (validation) results on diurnal variation / precision / biases! one-hand quality selection of the FTIR retrievals of all groups at IFU: 1. requested full FTIR data sets excluding the data quality selection usually performed (i.e., including all bad results) 2. applied homogenized selection thresholds for all FTIR data-sets: FTIR spectra quality (rms of the spectral fitting residuals) SZA min-max maximum number of iterations outlier-threshold (deviation from daily mean) all thresholds carefully tuned as a trade-off between reduction of scatter versus data loss Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 15 AV i ( i ) FTIR retrieval homogenization: data selection / quality control strongly impacts (validation) results on diurnal variation / precision / biases! i) start: raw results then: independent data selection per dofs; different result ensembles per dofs (real life) ii) start: one coincident ensemble for all dofs including data sdelction (strict selection!) then random generation of different sub- ensembles AV i ( i ) Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 16 consistent set of a priori profiles: harmonized tropopause adaptation consistent set of Tikhonov L 1 regularization matrices and altitude grids harmonized data selection / quality control FTIR retrieval homogenization: status July/August 2008 FTIR retrieval harmonization applied to 2003 & 2004 CH4 columns/profiles for 15 sites Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 17 Validation results: the 15 FTIR stations of our validation study stationlatitudelongitudestation altitude number of columns in 2003/2004 NCEP mean tropopause height Spitzbergen Thule* Kiruna Harestua Bremen Zugspitze Garmisch* Jungfraujoch Toronto* Izaa Paramaribo St-Denis Wollongong* Lauder* Arrival Heights* N N N N N N N N N N 5.81 N S S S S E W E E 8.85 E E E 7.99 E W W W E E E E 20 m 225 m 419 m 596 m 29 m 2964 m 745 m 3580 m 174 m 2367 m 7 m 50 m 40 m 370 m 200 m km 8.51 km 9.62 km km km km km km km km km km km 9.06 km *non-HYMN FTIR stations that asked me to participate in our validation paper Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 18 Harmonizing the g.-b. XCH4 validation data set: how to calculate the air column in situ p measurements not continuously available at all sites for 2003/2004 radio sondes not available for all sites NEP p-profiles and ECMWF in situ p available NCEP favored because: The forward model of the g.-b. FTIR retrievals uses NCEP pressure profiles to calculate the airmass-profile (fas.mas file) which is internally used to transfer VMR profiles to partial column (CH4) profiles I.e., the air columns calculated from NCEP are i) readily available, and ii) perfectly consistent to the CH4 total columns retrieved. Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 19 Harmonizing the g.-b. XCH4 validation data set: dry versus wet XCH4 previous SCIA validation studies derived XCH4 from FTIR via rationing by wet air columns (derived from pressure data). inconsistency to SCIAMACHY, which retrieves dry XCH4 (by rationing to CO2). FTIRs should substract the water vapor column from the air column: where do we get water columns from? not all FTIR groups have a maturated FTIR water retrieval; are NCEP water columns reliable? Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 20 Harmonizing the g.-b. XCH4 validation data set: how to calculate dry XCH4 validate NCEP water columns against maturated Zugspitze FTIR water vapor column retrieval (Sussmann, Borsdorff, Camy-Peyret, 2008) NCEP water ok, use for all stations Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 21 Harmonizing the g.-b. XCH4 validation data set: dry versus wet XCH4 impact on bias, annual cycle Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 22 Harmonizing the g.-b. XCH4 validation data set: different ray tracing algos e.g., at Zugspitze and Jungfraujoch . impact on annual cycles? No. Jungfraujoch FTIR Zugspitze FTIR daily means Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 23 Harmonizing the g.-b. XCH4 validation data set: altitude correction for high altitude FTIR sites Solution a (Dils, 06): use TM4 model ratio mountain/ground as correction factor for XCH4 Drawback: spatial resolution of the model (2 3 ) does not correspond with SCIAMACHY pixels Solution b (IFU, 08): add CH4 a priori column between surface and mountain top, scaled by retrieved scaling factor of the column above surface level SCIA pixel selection radius tropopause altitude mountain altitude FTIR Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 24 Validate IFU altitude correction: Zugspitze (2964 m) versus Garmisch (745 m) XCH4 uncorrected XCH4 after altitude correction to km Zugspitze (2.964 km), 2004 average E E-6 Garmisch (0.745 km), 2004 average E-6 difference1.8 %0.4 % derived from all 2004 coincidences Zuspitze/Garmisch Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 25 Validation strategy: spatial co-location criteria Problem: question of appropriate spatial co-location criteria - no unique answer Impact of selection radius on : number of coincides inclusion of remote sources/sinks magnitude of dynamically induced variability Solution: Use different collocation criteria in parallel (pixel selection radii 200 km, 500 km, 1000 km) and analyse their impact on the validation results [Sussmann et al. 2005] SCIA pixel selection radius tropopause altitude surface level FTIR location Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 26 Validation results: validation plots (daily means FTIR, SCIA-200, -500, km) Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 27 Validation results: validation plots (daily means FTIR, SCIA-200, -500, km) Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 28 Validation results: validation plots (daily means FTIR, SCIA-200, -500, km) Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 29 Validation results: validation plots (daily means FTIR, SCIA-200, -500, km) Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 30 Validation results: validation plots (daily means FTIR, SCIA-200, -500, km) Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 31 Validation results: validation plots (daily means FTIR, SCIA-200, -500, km) Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 32 Validation results: validation plots (daily means FTIR, SCIA-200, -500, km) Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 33 Validation results: validation plots (daily means FTIR, SCIA-200, -500, km) Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 34 Validation results: validation plots (daily means FTIR, SCIA-200, -500, km) Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 35 Validation results: validation plots (daily means FTIR, SCIA-200, -500, km) Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 36 Validation results: validation plots (daily means FTIR, SCIA-200, -500, km) Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 37 Validation results: validation plots (daily means FTIR, SCIA-200, -500, km) Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 38 Validation results: validation plots (daily means FTIR, SCIA-200, -500, km) Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 39 Table 3. Validation results. *: too few data / no significant results Frankenbergs Data Buchwitzs Data i : index of day AV i (n i ) : average number of FTIR spectra (or SCIA pixels) per day AV i ( i ) : diurnal variation of individual measurements (FTIR), pixel-to-pixel scatter (SCIA) within daily selection radius AV i ( i / n i ) : precision of daily mean Validation results: numbers on precisions, variabilities, and biases 60 Tables for: & 2004 data - 15 stations - 2 SCIA processors Buchwitzs Data Frankenbergs Data Zugspitze 2003 (dofs=1.91) AV i (n i ) AV i ( i ) [%] AV i ( i / n i ) [%] of daily means corrected for annual cycle [%] Bias [%] FTIR SCIA 200 km SCIA 500 km SCIA 1000 km Example Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 40 Validation results: number of SCIAMACHY data Spitzbergen 2004 (dofs=2.22) AV i (n i ) FTIR2.00 SCIA 200 km SCIA 500 km SCIA 1000 km Buchwitzs Data Zugspitze 2003 (dofs=1.91) AV i (n i ) FTIR5.29 SCIA 200 km SCIA 500 km SCIA 1000 km Frankenbergs Data i : index of day AV i (n i ) : average number of FTIR spectra per day / of SCIA pixels within a certain selection radius per day Results for SCIA AV i (n i ) : i) more data from IMAP compared to WFMD ii) extremely few data for WFMD at the poles iii) numbers due to differing data selection procedures Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 41 Validation results: single column measurement precision FTIR (I) i : index of day AV i ( i ) : diurnal variation of individual measurements (FTIR) different numbers due to differing data selection procedures XCH4 i of day i (18 Sep) = 0.13 % n i = 9 columns, 10 min integration per column diurnal variation mainly due to clouds Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 42 Validation results: single columns precision FTIR (II) Results for FTIR AV i ( i ) : i) FTIR single measurement precision is typically around 0.5 % for most sites. It is interesting to compare the value for Zugspitze 2003 data (0.56 %) with the value of 1.3 % obtained from our older version FTIR retrievals for 2003, see Table 1 in [Sussmann et al., 2005]: Improved FTIR precision is due to the optimized FTIR profile retrieval (Tikhonov harmonizaton) compared to the simple VMR profile scaling performed in ii) There are sites with slightly better-than-average precision ( %), e.g., Jungfraujoch and Izana, which points to a very high engineering quality of the spectra and/or measurement conditions with very low cloud contamination. iii) There are also sites with slightly reduced precision (1-2 %), e.g., Thule and Wollongong, which might be due, e.g., to high cloudiness. different numbers due to differing data selection procedures Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 43 Results for SCIA AV i ( i ) : i) Pixel-to-pixel variability is typically in the order of %. Compare WFM-DOAS v1.0 for Zugspitze 2003 within 1000 km radius (1.60 %) with the value of 5.4 % for same radius but WFM-DOAS v0.41 [Sussmann et al., 2005]: This reflects a significant improvement in SCIA precision. ii) Usually slight increase of AV i ( i ) with the selection radius (e.g., Lauder IMAP-DOAS for 2003: 1.6 %, 2.01 %, 2.24 % for 200 km, 500 km,1000 km) sources/sinks included. iii) Slight decrease of AV i ( i ) with increasing selection radius for island sites (e.g., Izana and St. Denis) no soures above ocean but homogeneous ocean OH sink added. iv) Pixel-to-pixel variability typically smaller for WFM-DOAS v1.0 by 0.5 % compared to IMAP-DOAS v49 (in line with WFMD stronger quality control, at the cost of fewer data). Validation results: single columns precision SCIA (I) i : index of day AV i ( i ) : pixel-to-pixel variability of all SCIA columns of a day within a certain selection radius 200 km Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 44 Validation results: response to day-to-day variability, 2 definitions indiv. columns Remember AV i ( i ) = precision of a single column measurement I) Therefore AV i ( i / n i ) = stat. error (precision) of a daily mean (FTIR or SCIA) II) day-to-day variability = of daily means corrected for annual cycle [%] : AV i ( i /sqrt (n i )) has to be smaller than the true day-today variability in order to obtain reasonable results. SCIA 200 km daily means FTIR Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 45 Validation results: response to day-to-day variability, FTIR Results FTIR i) The precision of daily mean XCH4 AV i ( i /sqrt (n i )) at different stations is 0.15 % % with values up to 1 %, depending on the single measurement precision and the number of FTIR measurements per day. ii) The day-to-day variability ( of daily means corrected for annual cycle [%]) observed by FTIR is typically %. These numbers should be a approximation of true XCH4 day-to-day variability above the sites because the precision of the daily mean is typically a significantly smaller number, see i). Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 46 Validation results: response to day-to-day variability, SCIA (I) i) AV i ( i /sqrt (n i )) for is typically < 0.5 % for the 200 km selection radius and is of course decreasing with higher selection radii. This means that the true day-to-day variability of % (as seen by FTIR) can be measured by IMAP-DOAS v1.0. iii) For the 200 km radius the limited precision of the daily means tends to slightly impact the day to-day-variability towards slightly too high values. ii) For the 500 km radius the day-to-day variability derived from IMAP-DOAS v49 typically agrees to the value found from FTIR. Results IMAP-DOAS v49 Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 47 Validation results: response to day-to-day variability, SCIA (II) iv) For the 1000 km radius the SCIA day-to-day variability is slightly reduced relative to what is seen by FTIR above a single site. This is exactly what we expect due to an averaging of the impact of planetary wave activity (driver of day-to-day variability) within large selection radii: SCIA pixel selection radius tropopause altitude surface level FTIR location continued : Results IMAP-DOAS v49 Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 48 Validation results: response to day-to-day variability, SCIA (III) i) The precision of the daily mean AV i ( i /sqrt (n i )) is lower for WFM-DOAS v1.0 compared to IMAP-DOAS v1.0 due to the smaller number of data n i. True day-to-day variability (as seen by FTIR) can be reflected approximately, but there is a tendency for over-estimation due to the limited precision of the daily means. Result WFM-DOAS v1.0 Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 49 Validation results: biases, latitudinal dependency (I) without altitude correction Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 50 Validation results: biases, latitudinal dependency (II) with altitude correction Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 51 Validation results: biases, latitudinal dependency (III) with altitude correction Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 52 Results i) the biases for Thule (76.53 N) and Toronto (43.66 N) are less reliable because there are strong differences between the years 2003 and 2004 and there are also strong differences for the different selection radii ii) there is an overall high bias around % for SCIAMACHY relative to FTIR iii) bias is slightly higher (by 0.5 %) for WFM-DOAS v1.0 compared to IMAP-DOAS v49 iv) there is no significant latitudinal dependency of the bias (but high Paramaribo FTIR XCH4: real?). Validation results: biases, latitudinal dependency (IV) Research Center KarlsruheRalf Sussmann et al. IMK-IFU Garmisch-PartenkirchenHarmonized global SCIA XCH4 validation 53 Single pixel column measurement precision of SCIAMACHY XCH4 is % and has improved by a factor 2-4 since 2005 (rel. to WFM-DOAS v0.41). WFM-DOAS v1.0 shows smaller pixel-to-pixel variability (by 0.5 %) due to stronger data selection at the cost of fewer data compared to IMAP-DOAS v49 Improved precision allows now IMAP-DOAS v49 to retrieve via a relatively small 200-km pixel selection radius realistic values ( %) on the day-to-day variability of XCH4. due to the smaller number of data for WFM-DOAS v1.0 the precision of the daily means is smaller and the day-to-day variability tends to be over-estimated Still no significant correlation of the annual cycles between SCIA and FTIR can be found (frown-shaped (SCIA) versus -sine or smiley shape (FTIR)). There is a high bias of SCIAMACHY (IMAP-DOAS v49 and WFM-DOAS v1.0) of % relative to a strongly improved and harmonized g.-b. FTIR data set comprising 15 globally distributed stations. Bias is slightly higher (by 0.5 %) for WFM-DOAS v1.0 compared to IMAP-DOAS v49 There is no significant latitudinal dependency of the bias (high for Paramaribo). Conclusions