Search

River channels store large volumes of water globally, critically impacting ecological and biogeochemical processes. Despite the importance of river channel storage, there is not yet an observational constraint on this quantity. We introduce a 26‐year record of entirely remotely sensed volumetric channel water storage (CWS) change on 26 major world rivers. We find mainstem volumetric CWS climatology amplitude (CA) represents an appreciable amount of basin‐wide terrestrial water storage variability (median 2.78%, range 0.04%–12.54% across world rivers), despite mainstem rivers themselves represent an average of just 0.2% of basin area. We find that two global river routing schemes coupled with land surface models reasonably approximate CA (within ±50%) in only 11.5% (CaMa‐Flood) and 30.7% (HyMap) of rivers considered. These findings demonstrate volumetric CWS is a useful quantity for assessing global hydrological model performance, and for advancing understanding of spatial patterns in global hydrology. Plain Language Summary: Rivers are a critical part of global hydrology, but until now the variation in how much water rivers store has not been observed directly (via a measurement based approach) on the global scale. Surface water storage and fluxes are critical in understanding the impacts and trajectory of global climate change. We created a 26 years record of this quantity across 26 of the world's largest rivers. We found that the storage variation in river main channels can represent up to 12.54% of the total water storage variation in a river basin despite only representing 0.2% of the total surface area. We also find that the only previous method to estimate this quantity, through modeling (global river routing schemes coupled with land surface models), represent this quantity within 50% of our estimated value on between just 11.5%–30.7% of the rivers we studied. Our work shows that this quantity is a substantial land surface hydrology storage term, and also demonstrates a critical need for the inclusion of more measurement based data to improve model performance. Global models are our best tool for understanding global climate change and data sets like this one can help improve their accuracy. Key Points: We introduce a 26‐year record of entirely remotely sensed volumetric channel water storage anomalyStorage climatology amplitude represents (0.04%–12.54%) terrestrial water storage variability but just 0.2% of basin areaThis new quantity provides a physical grounding for land surface hydrology models that are critical for the water cycle [ABSTRACT FROM AUTHOR]

Ocean wind speed and/or wind vector measurements from the cyclone global navigation satellite system (CYGNSS), advanced scatterometer (ASCAT), and WindSat satellites are used to investigate diurnal wind variations over the tropical ocean using a method previously reported for diurnal amplitude estimation. The resulting diurnal variability in wind speed is examined in terms of its spatial and temporal properties. The results also show that the remotely sensed diurnal amplitudes obtained are largely consistent with those predicted by the existing reanalysis products such as the European Centre for Medium-range Weather Forecasts (ECMWF) ERA5. [ABSTRACT FROM AUTHOR]

An algorithm for detecting coherence in Cyclone Global Navigation Satellite System (CYGNSS) mission delay-Doppler maps (DDMs) is presented. Because CYGNSS DDMs report only the observed power without phase information, the algorithm uses estimates of power “spread” within the DDM to flag coherency. Since the estimate used is a ratio of the powers in differing portions of the DDM, it is less sensitive to absolute power calibration and to the GPS C/A code type observed, and is applied to CYGNSS Level-1 uncalibrated DDMs. The basic detector formulation is described along with modifications to improve performance in lower signal-to-noise ratio (SNR) situations. The required detection thresholds are determined using matchups with CYGNSS “Raw I/F” mode measurements for which the DDM phase can be computed and used to identify coherence more precisely. Application of the final detector over a large CYGNSS data set suggests that approximately 8.9% of all inland returns are coherent. Inland regions persistently identified as coherent were found largely to be associated with the presence of water bodies. A smaller set of desert locations apparently having very low surface roughness were also found to be associated with persistent coherence. The detector was also applied to a set of ocean measurements, with the results showing that persistent coherence is limited to areas with sheltered waters. Ocean tests avoiding such regions indicate that the detector’s false-alarm rate is approximately 0.0012% for the detection threshold used. [ABSTRACT FROM AUTHOR]

The capabilities of radar altimetry to measure inland water bodies are well established, and several river altimetry datasets are available. Here we produced a globally distributed dataset, the Global River Radar Altimeter Time Series (GRRATS), using Envisat and Ocean Surface Topography Mission (OSTM)/Jason-2 radar altimeter data spanning the time period 2002–2016. We developed a method that runs unsupervised, without requiring parameterization at the measurement location, dubbed virtual station (VS) level, and applied it to all altimeter crossings of ocean-draining rivers with widths >900 m (>34 % of the global drainage area). We evaluated every VS, either quantitatively for VS locations where in situ gages are available or qualitatively using a grade system. We processed nearly 1.5 million altimeter measurements from 1478 VSs. After quality control, the final product contained 810 403 measurements distributed over 932 VSs located on 39 rivers. Available in situ data allowed quantitative evaluation of 389 VSs on 12 rivers. The median standard deviation of river elevation error is 0.93 m, Nash–Sutcliffe efficiency is 0.75, and correlation coefficient is 0.9. GRRATS is a consistent, well-documented dataset with a user-friendly data visualization portal, freely available for use by the global scientific community. Data are available at 10.5067/PSGRA-SA2V1 (Coss et al., 2016). [ABSTRACT FROM AUTHOR]

Satellite altimetry has been proven as an effective technology to accurately measure water level, ice elevation, and flat land surface changes since the 1990s. To overcome limitations of pulse-limited altimetry, new altimetric missions such as Cryosat-2 and Satellite with ARgos and AltiKa (SARAL/AltiKa), have been designed to have higher along-track spatial resolution to measure more accurately inland water levels for small water bodies, and coastal sea level changes. In this study, we evaluate the performance of Cryosat-2 low-resolution (LRM) and SARin modes and SARAL/AltiKa Ka-band data on two connected lakes in central Tibetan Plateau, and in the coastal region of Taiwan. Results are compared with in situ tide gauge data in Taiwan and altimetric lake level time series from the CNES Hydroweb database. Our results show that water level change trends observed by Cryosat-2 20-Hz retracked observations, the SARAL/AltiKa 40-Hz Ice-1 retracked data, and the Hydroweb measurements are consistent with the estimated water level trend of ∼0.30 m/y, during 2011–2017, and 2013–2015, for the Tibetan Migriggyangzham Co and Dorsoidong Co, respectively. For the coastal region, the performance of SARAL/AltiKa is better than that of Cryosat-2 LRM data in Taiwan. This finding demonstrates the superiority of the Ka-band over Ku-band radar altimetry. [ABSTRACT FROM AUTHOR]

Due to solar heating, many geophysical quantities, such as precipitation, sea surface temperature, and wind, have distinct diurnal signatures in the tropical region. To explore the potential of wind speed measurements of the Cyclone Global Navigation Satellite System (CYGNSS) mission launched in December 2016, we examine the degree to which wind diurnal cycles can be estimated from the surface wind speed data provided by CYGNSS along with data from the ASCAT mission. Ocean wind speed measurements are simulated based on the Cross-Calibrated Multi-Platform analysis fields of surface wind vectors. We estimate the amplitude and phase of diurnal wind speed based on a discrete Fourier transform generalized to analyze unevenly sampled time series of wind speed anomalies. The results show that the combination of CYGNSS and ASCAT measurements can provide some ability to observe diurnal wind information, but that additional time samples (e.g., from other measurement sources) would likely provide further benefits in estimation performance. [ABSTRACT FROM AUTHOR]

The Cyclone Global Navigation Satellite System (CYGNSS) is a spaceborne GNSS-reflectometry mission, which was launched on December 15, 2016 for ocean surface wind speed measurement. CYGNSS includes eight small satellites in the same low earth orbit, so that the mission provides wind speed products having unprecedented coverage both in time and space to study multitemporal behaviors of oceanic winds. The nature of CYGNSS coverage results in some locations on earth experiencing multiple wind speed measurements within a short period of time (a “clump” of observations in time) resulting in a “rapid revisit” series of measurements. Such observations seemingly can provide indications of regions experiencing rapid changes in wind speeds, and therefore serve as an indicator of convective activity. An initial investigation of this concept using simulated and on-orbit CYGNSS measurements is provided in this paper. The temporally “clumped” properties of CYGNSS measurements are examined, and the results show that clump durations and spacing vary with latitude. For example, the duration of a clump can extend as long as a few hours at higher latitudes, with gaps between clumps ranging from 6 to as high as 12 h depending on latitude. Initial examples are provided to indicate the potential of changes within a clump to detect convective activity through a comparison with convective activity indicators derived from model datasets. The results at present are limited by the ongoing calibration of CYGNSS wind speed retrievals, so that future work will be required to obtain a more complete assessment, but nevertheless clearly indicate the potential utility of the method for studies of atmospheric convection. [ABSTRACT FROM AUTHOR]

Repeat-track analysis is commonly utilized to generate elevation change time series from satellite radar altimetry over ice sheets. It requires surface gradient (SG) correction due primarily to orbital drifts and radar-related empirical corrections caused by radar scatters from ice surface and potential subsurface. In this letter, two approaches, namely, the use of a digital elevation model (DEM) and the modified repeat-track analysis, which uses the accumulated Envisat altimetry profiles, are applied to correct the SG over both Greenland ice sheet (GrIS) and Antarctic ice sheet (AIS). By comparing the root mean square (rms) of elevation change time series after SG correction, the percentage of data (rms< 1 m) obtained by using modified repeat-track analysis is found to be 85% and 88% for the GrIS and AIS, respectively, as opposed to 45% and 44% if the DEM method is used. Furthermore, three cases are studied to assess empirical corrections for elevation retrieved from both ice-1 and ice-2 algorithms over the AIS. We conclude that the modified repeat-track analysis is more effective to remove topographic induced error. For the ice-2 algorithm, waveform shape parameters are needed in addition to applying corrections from changes in backscatter coefficients. The trend of elevation changes from the ice-1 algorithm with only backscatter analysis agrees with that from the ice-2 algorithm with corrections from backscatter coefficient changes and waveform shape parameters. This study could provide a potential data processing recipe for generating improved satellite radar altimetry elevation time series over ice sheets. [ABSTRACT FROM PUBLISHER]

The Ganges-Brahmaputra-Meghna (GBM) river basins exhibit extremes in surface water availability at seasonal to annual time scales. However, because of a lack of basinwide hydrological data from in situ platforms, whether they are real time or historical, water management has been quite challenging for the 630 million inhabitants. Under such circumstances, a large-scale and spatially distributed hydrological model, forced with more widely available satellite meteorological data, can be useful for generating high resolution basinwide hydrological state variable data [streamflow, runoff, and evapotranspiration (ET)] and for decision making on water management. The Variable Infiltration Capacity (VIC) hydrological model was therefore set up for the entire GBM basin at spatial scales ranging from 12.5 to 25 km to generate daily fluxes of surface water availability (runoff and streamflow). Results indicate that, with the selection of representative gridcell size and application of correction factors to evapotranspiration calculation, it is possible to significantly improve streamflow simulation and overcome some of the insufficient sampling and data quality issues in the ungauged basins. Assessment of skill of satellite precipitation forcing datasets revealed that the Tropical Rainfall Measuring Mission (TRMM) Multisatellite Precipitation Analysis (TMPA) product of 3B42RT fared comparatively better than the Climate Prediction Center (CPC) morphing technique (CMORPH) product for simulation of streamflow. The general conclusion that emerges from this study is that spatially distributed hydrologic modeling for water management is feasible for the GBM basins under the scenario of inadequate in situ data availability. Satellite precipitation forcing datasets provide the necessary skill for water balance studies at interannual and interseasonal scales. However, further improvement in skill may be required if these datasets are to be used for flood management at daily to weekly time scales and within a data assimilation framework. [ABSTRACT FROM AUTHOR]

Two methods to assess ocean tide models, the current method and the total discrepancy method, are compared from the perspective of their relationship to the root-mean-square difference of tidal sea surface height (total discrepancy). These two methods are identically the same when there is only one spatial location involved. When there is more than one spatial location involved, the current method is the root-mean-square difference of total discrepancy at each location, and the total discrepancy method is the averaged total discrepancy. The result from the current method is always larger than or equal to that from the total discrepancy method. Monte Carlo simulation indicates that the difference between their results increases with increasing spatial variability of total discrepancy. Both of these two methods are then used to compare the two tide models of the Ocean Surface Topography Mission (OSTM)/ Jason-2. The discrepancy of these two models as measured by the total discrepancy method decreases monotonically from around 11.4 to 2.2 cm with depth increasing from 50 to 1000 m. In contrast, the discrepancy measured by the current method varies from 21.6 to 2.9 cm. Though the discrepancy measured by the current method decreases with increasing depth in general, there are abrupt increases at several depth ranges. These increases are associated with large spatial variability of total discrepancy and their physical explanation is elusive. Because the total discrepancy method is consistent with the root-mean-square difference of tidal sea surface height and its interpretation is straightforward, its usage is suggested. [ABSTRACT FROM AUTHOR]

We explored the application of satellite radar altimetry for the monitoring of small inland bodies of water and hydrologic studies using a water-detection algorithm, optimally retracked TOPEX/POSEIDON data at 10-Hz sampling, and investigated the use of radar backscatter to improve land cover classification. The procedure was demonstrated over Manitoba and south-western (SW) Ontario, and the Amazon River Basin study regions. Compared with an L-band synthetic aperture radar data generated water-land cover mask, the water-detection algorithm detected more water points over the Amazon basin. High correlation of 0.98 between the retracked 10-Hz altimetry and the gauge measurements in Manitoba confirmed that the retracked TOPEX data are more accurate than the non-retracked data, and with higher along-track spatial resolution by virtue of its higher sampling at 10 Hz. [ABSTRACT FROM AUTHOR]

This study presents results of Jason-1 (J1) and Jason-2 (J2) radar altimetry absolute calibration (cal/val) over the Marblehead lake water level gauge, with an aid of a GPS buoy, in Lake Erie, the Great Lakes, in North America. The altimeter bias is estimated using the height difference between the altimeter lake surface height and the in situ data in each altimeter 10-day repeat cycle. The altimeter bias estimates for J1 are 81 ± 2 and 70 ± 2 mm for Geophysical Data Record (GDR) Versions 'B' and 'C', respectively, and 148 ± 5 mm (GDR) and 147 ± 7 mm (IGDR) for J2, respectively. The bias estimates are slightly smaller compared with estimates at other dedicated calibration sites such as the Harvest Platform, the Corsica Site, and the Bass Strait site due in part to the effect of the sea state bias. The J2-to-J1 relative bias determined in the tandem mode over Lake Erie is 84 ± 28 mm, which agrees well with the results from the global analysis and the dedicated sites aforementioned. [ABSTRACT FROM AUTHOR]

In this study, we validated the Jason-2 (J2) dual-frequency ionosphere delay measurements in terms of vertical total electron content (VTEC) in the Geophysical Data Record (GDR) with the coarse resolution JPL Global Ionospheric Maps (GIM), the regional ionospheric maps generated by the Crustal Motion Observation Network of China (CMONOC), and the Jason-1 (J1) interleaved tandem mission (with J2) ionosphere delay measurements. The estimates of the relative biases and their uncertainties (95% confidence, in TEC units) for various comparison cases with different data spans and regions are (1) J2-GIM: -3.07 ± 0.18 TECu (or 6.75 ± 0.40 mm in the range delay), (2) J2-CMONOC: -2.87 ± 0.38 TECu (6.31 ± 0.84 mm), (3) J1-GIM: -0.19 ± 0.18 TECu (0.42 ± 0.40 mm), and (4) J2-J1 via double-differencing, that is, (J2-GIM)-(J1-GIM): -2.88 ± 0.26 TECu (6.34 ± 0.57 mm). The scatter (RMS, about the mean differences) between J2 and GIM, CMONOC, or J1 is about 3∼8 TECu (6.6∼17.6 mm). We conclude that the results from global and regional analysis are consistent and that the J2 ionosphere delay is about 6∼10 mm (95% confidence) shorter than the delay computed by GIM, J1, or the regional model. [ABSTRACT FROM AUTHOR]

It has been more than a decade since the last comprehensive accuracy assessment of global ocean tide models. Here, we conduct an evaluation of the barotropic ocean tide corrections, which were computed using FES2004 and GOT00.2, and other models on the Jason-2 altimetry Geophysical Data Record (GDR), with a focus on selected coastal regions with energetic ocean dynamics. We compared nine historical and contemporary ocean tide models with pelagic tidal constants and with multiple satellite altimetry mission (T/P, ERS-1/-2, Envisat, GFO, Jason-1/-2) sea level anomalies using variance reduction studies. All accuracy assessment methods show consistent results. We conclude that all the contemporary ocean tide models evaluated have similar performance in the selected coastal regions. However, their accuracies are region-dependent and overall are significantly worse than those in the deep-ocean, which are at the 2-3 cm RMS (root-mean-square) level. The Gulf of Mexico and Northwest Atlantic regions present the least reduction of altimetry sea surface height variability after ocean tides are removed, primarily because of large oceanic variability associated with loop currents in the Gulf of Mexico and the Gulf Stream in the Northwest Atlantic. [ABSTRACT FROM AUTHOR]

We validated Jason-2 satellite altimeter Sensor Geophysical Data Records (SGDR) by retracking 20-Hz radar waveforms over the California coastal ocean using cycles 7-34, corresponding to September 2008-June 2009. The performance of the ocean, ice, threshold, and modified threshold retrackers are examined using a reference geoid based on Earth Gravitational Model 2008 (EGM08). Over the shallow ocean (depth < 200 m), the modified threshold retracker, which is developed for noisy waveforms with preleading edge bump, outperforms the other retrackers. It is also shown that retracking can improve the precision of sea surface heights (SSHs) for areas beyond 2-5 km from the shore. Although the ocean retracker generally performs well over the deep ocean (depth > 200 m), the ocean-retracked SSHs from some of the cycles are found to be less precise when the waveforms do not conform to the Brown ocean model. We found that the retrackers developed for nonocean surfaces can improve the noisy ocean-retracked SSHs. Among the retrackers tested here, the ice retracker overall provides the most precise SSH estimates over the deep ocean in average using cycles 7-34 in the study region. [ABSTRACT FROM AUTHOR]

Previous studies using satellite radar altimetry to observe inland river and wetland water level changes usually spatially average high-rate (10-Hz for TOPEX, 18-Hz for Envisat) measurements. Here we develop a technique to apply retracking of TOPEX waveforms by optimizing the estimated retracked gate positions using the Offset Center of Gravity retracker. This study, for the first time, utilizes stacking of retracked TOPEX data over Louisiana wetland and concludes that the water level observed by each of 10-Hz data with along-track sampling of ∼660 m exhibit variations, indicating detection of wetland dynamics. After further validations using nearby river gauges, we conclude that TOPEX is capable of measuring accurate water level changes beneath heavy-vegetation canopy region (swamp forest), and that it revealed wetland dynamic flow characteristics along track with spatial scale of 660 m or longer. [ABSTRACT FROM AUTHOR]

Abstract: Tides in the polar region play a major role in the dynamics of sea ice and floating glacial ice shelves. Existing ocean tide models are much less accurate in coastal and shallow seas and polar oceans than other ocean areas. In particular, ocean tides are largely unknown in parts of polar oceans that are covered by permanent or seasonal sea ice and in regions that are beyond the coverage of the TOPEX/POSEIDON (T/P). We conducted a periodogram analysis using simulated satellite altimeter data (T/P or JASON, ERS-2 or ENVISAT) at single and dual satellite crossover locations. Our results indicate that the TOPEX/ERS-2 dual satellite crossovers show considerable improvements to mitigate tidal aliasing and are, therefore, preferable for tidal modeling. Empirical ocean tide models have been determined using the T/P and ERS-2 altimetry at crossovers in the Southern Ocean below 55°S and employing the response method. Evaluations using in situ pelagic tidal constants and altimeter data indicate that the resulting empirical tide models are as good as the selected contemporary models (NAO.99b, TPXO.6.2, and CATS02.01) in the Southern Ocean. Solutions for the M2, O1, N2, and Q1 tides that are obtained from the ERS-2 data only show a reasonable accuracy, indicating that extensions of the tide modeling domain beyond the T/P coverage are feasible. [Copyright &y& Elsevier]

We conducted an assessment of the TOPEX dual-frequency nadir ionosphere observations in the TOPEX/Poseidon (T/P) GDR by comparing TOPEX with the Center for Orbit Determination in Europe (CODE) Global ionophore Map (GIM), the climatological model IRI2001, and the DORIS (onboard T/P) relative ionosphere delays. We investigated the TOPEX (TOPEX Side A and TOPEX Side B altimeters, TSA and TSB, respectively) ionosphere observations for the time period 1995-2001, covering periods of low, intermediate, and high solar activity. Here, we use absolute path delays (at Ku-band frequency of the TOPEX altimeter and with positive signs) rather than Total Electron Content (TEC). We found significant biases between GIM and TOPEX (GIM-TOPEX) nadir ionosphere path delays: -8.1 ± 0.4 mm formal uncertainties and equivalent to 3.7 TECu) and -9.0 ± 0.7 mm (4.1 TECu) for TSA and TSB, respectively, indicating that the TOPEX path delay is longer (or with higher TECu) that GIM. The estimated relative biases vary with latitude and with daytime or nighttime passes. The estimated biases in the path delays (DORIS-TOPEX) are: -10.9 ± 0.4 mm (5.0 TECu) and -14.8 ± 0.6 mm (6.7 TECu), for TSA and TSA, respectively. There is a distinct jump of the DORIS path delays (-3.9 ± 0.7 mm, TSA delays longer than TSB delays) at the TSB altimeter switch in February 1990, presumably due to inconsistent DORIS processing. The origin of the bias between GIM (GPS, L-band) and TOPEX (radar altimeter, Ku-band) is currently unknown and warrants further investigation. Finally, the estimated drift rates between GIM and TSA, DORIS and TSA ionosphere path delays for the 6-year study span are -0.4 mm/yr and -0.8 mm/yr, respectively, providing a possible error bound for the TOPEX/Poseidon sea level observations during periods of low and intermediate solar activity. [ABSTRACT FROM AUTHOR]

Abstract: A new method to estimate the vertical crustal motion from satellite altimetry over land was developed. The method was tested around Hudson Bay, where the observed vertical motion is largely caused by the incomplete glacial isostatic adjustment (GIA) as a result of the Laurentide ice sheet deglaciation since the last glacial maximum (LGM). Decadal (1992–2003) TOPEX/POSEIDON radar altimetry data over land surfaces were used. The results presented here are improved compared to a previous study (Lee, H., Shum, C.K., Kuo, C.Y., Yi, Y., Braun, A., 2008. Application of TOPEX altimetry for solid Earth deformation studies. Terr. Atmos. Ocean. Sci. 19, 37–46. doi:10.3319/TAO.2008.19.1-2.37(SA).) which estimated vertical motion only over relatively flat land surfaces (standard deviation of the height variation <40cm). In this study, we extended the concept of traditional 1-Hz (one-per-frame) radar altimeter ocean stackfiles to build 10-Hz (10-per-frame) land stackfiles over Hudson Bay land regions, and succeeded in obtaining vertical motion estimates over much rougher surfaces (standard deviation of the height variation <2m). 90-m C-band Shuttle Radar Topography Mission (SRTM) Digital Elevation Model (DEM) is used as a reference surface to select an optimal waveform retracker, to correct surface gradient errors, and to calculate land surface anomalies. Here, we developed an alternative retracker, called the modified threshold retracker, resulting in decadal vertical motion time series over a 1500km by 1000km region covering northern Ontario, northeastern Manitoba, and the Great Lakes region which is at the margin of the former Laurentide ice sheet. The average of the estimated uncertainties for the vertical motion is 2.9mm/year which is comparable to 2.1mm/year of recent GPS solutions. The estimated vertical motion is compared with other geodetic observations from GPS, tide gauge/altimetry, GRACE, and several GIA models. The data agree best with the laterally varying 3D GIA model, RF3S20 (β =0.4) whereas the combination of land altimetry solution with other measurements match best with the models RF3S20 (β =0.0) or RF3S20 (β =0.2) in terms of mean and standard deviation of the differences. It is anticipated that this innovative technique could potentially be used to provide additional constraints for GIA model improvement, and be applied to other geodynamics studies. [Copyright &y& Elsevier]