Since ocean bottom pressure represents a column integral of the mass of the atmosphere plus ocean, this measurement technique permits the deduction of ocean bottom pressure changes from space. Germany provides also the Eurockot launch vehicle. This material, with a very low coefficient of thermal expansion, provides the dimensional stability necessary for precise range change measurements between the two spacecraft.
The sensors include: 4. The actuators include a cold gas system with 12 attitude control thrusters and two orbit control thrusters, each rated at 40 mN and three magnetorquers.
Following separation, the leading GRACE satellite began pulling away from the trailing satellite at a relative speed of about 0. The two satellites in tandem formation are loosely controlled, they are separated at distances between to km apart. The onboard cold-gas propulsion system is being used to maintain the separation between km and km. Since mission launch, orbit maneuvers have been needed about every 50 days to do this. The spacecraft orbits have a 30 day repeat cycle, and a new gravity field is determined each month.
The GRACE system accuracy is sufficient to determine a change in mass equivalent to a volume of water with depth 1 cm over a radius of about km. A backup zenith receive antennae and a backup nadir transmit antenna SZA-Tx , along with the appropriate RF electronics assembly, complete the telemetry and telecommand subsystem.
CCSDS protocols are used for all data communication. The S-band frequencies for the two satellite system are:. The polar location of Svalbard makes it possible to have access to the data on almost all orbits. They represent changes in water storage related to weather, climate, and seasonal patterns. In , the continental United States suffered through one of its worst droughts in decades. As another summer arrives in North America, surface water conditions have improved in many places, but drought has persisted or deepened in others.
Underground, the path out of drought is much slower. The top map of Figure 6 shows the "wetness" or moisture content in the "root zone"? The bottom map of Figure 6 shows water storage in shallow aquifers. The current water content is compared to a long-term average for early June between and To see the monthly changes from August through May , download the animation of Ref.
The root zone map offers perspective on the short-term weeks to months water situation; for instance, the passage of a tropical storm can have a distinct impact on root zone moisture.
Compared to the summer of , moisture near the surface in June is significantly better in most of the eastern and northern portions of the continental United States, particularly the Midwestern areas around the Mississippi River. Flooding has instead become the problem in Montana and North Dakota. Portions of Arizona, Nevada, and southeastern California are extremely dry, even by desert standards. The bottom map of Figure 6 tells more of a long-range story. Groundwater takes months to seep down and recharge aquifers, and that clearly has not happened in the Rocky Mountain states and most of Texas.
Underground storage has improved in much of the southeastern and central U. Southern California has a deficit despite promising signs in the winter and spring. The GRACE mission has experienced battery degradation that requires careful electrical load and battery charging management.
This encompasses such obvious measures as the minimization of fuel usage and thruster cycles, but also the continuous optimization of parameter settings and the balancing of several consumables. Close interaction between the science- and operation- teams is required throughout, because the satellites themselves are part of the experiment. The resources on both GRACE satellites are still sufficient to prolong the mission until at least This article is organized as follows: in the following section we describe the time-variable Swarm gravity fields alongside an error assessment that we obtain from projecting monthly Swarm-only gravity fields 27 onto the spatial modes from GRACE.
We consider two variants of the reconstruction approach: in the first one, the entire observed mass change signal is decomposed and projected; in the second variant, a deterministic signal model is first estimated from the GRACE data, then removed prior to decomposition and ingesting Swarm data, before subsequently being restored. This is followed by a conclusion and the explanation of our methods. We have reconstructed monthly Swarm geopotential solutions and surface mass change maps measured in e.
As an example, Fig. The low-degree solution shows that Swarm-only gravity recovery Fig. The Swarm reconstruction Fig. March e.
The maps in Fig. Errors are below 10 cm for most regions. Largest RMS errors are found in the Amazon basin, where the large hydrological signal can neither be fully captured by Swarm nor fully represented in three modes. The degree and order 12 and 40 error maps appear broadly similar, with larger errors for higher degrees.
The reconstruction method reduces the errors significantly 0. The large errors observed at the start of the Swarm mission can be attributed mainly to higher ionospheric activity during and This affected the performance of Swarm GNSS receivers 37 , an effect which was mitigated through receiver updates in the early mission phase The reconstruction is much less affected by ionospheric disturbances as compared to the monthly Swarm-only solutions.
In order to assess the impact of the new approach for typical regional applications, we here compare area-mean mass changes in terms of e. Figure 3 shows basin averages for the locations, see Fig. Basin averages for different regions expressed in meter of e. The error of the reconstruction is indicated in yellow as computed by Eq. Study regions.
For each basin, 4 GPS stations were randomly chosen and analyzed. Remaining stations are included in the supplementary materials. Stations shown at the charts are marked. The first observation from Tables 1 and 2 is that, during the Swarm time frame considered here, the monthly GRACE-derived mass anomalies reveal a quite regular behaviour for all regions considered, without larger e.
ENSO-related temporal anomalies and without trend changes or similar obvious interannual changes. It is thus not surprising that this simple GRACE-based prediction model already provides a good fit, however this is obviously due to the low variability in the considered time frame and therefore cannot invalidate methods that rely on other data such as Swarm.
Trends derived from a 4-year period cannot provide significance for assessing geophysical processes, but they can be compared to each other. We notice, e. Again, this is due to only very moderate interannual variability in this timeframe. We notice that trends from a Swarm-only monthly solution are generally larger as compared to the reconstruction approach, which is in line with the methodology.
Results are diverse for the smaller regions. This is encouraging, but it may be expected due to the mentioned regular behaviour of mass change during the time frame. This affects the Antarctic regions in particular, as can be seen in the time series as well as in Figs. For this reason, we only consider the time from to for the Antarctic ice sheet, except if stated otherwise.
These rates are predominately driven by melting glaciers in West Antarctica and the Antarctic peninsula, while the East Antarctic ice sheet appears more stable. As expected, the monthly Swarm-only solutions suffer from considerable noise, in particular in the beginning. For the Amazon basin, GRACE exhibits a large seasonal signal overlaid by a declining trend in the years and , which reverses in The noise of the monthly Swarm-only solutions 0.
In general, the signal-to-noise ratio in Greenland is quite low, as can be seen in Fig. The Ganges basin is the smallest study region considered here approx. For example, Ref. On the basis of comparisons to GRACE-derived basin averages, our findings are that monthly Swarm reconstructions should generally be preferred to a simple six-parameter GRACE model constant, trend, annual and semiannual terms , although we admit that the simple model, in certain situations, is sufficient.
A detailed analysis can be found in Section S4 of the Supplement. In order to validate the new approach with independent data, we analyze time series from twenty globally distributed GPS Global Positioning System sites. Vertical GPS displacements were pre-processed and corrected for non-tidal atmospheric, non-tidal oceanic and post-glacial rebound effects, as described in the Data section of the Supplement.
We find that both Swarm-reconstructed and GRACE-predicted displacements reproduce well inter-annual signals observed by GPS, while the fits for trends and at the annual timescale are moderate. Similarly we find a good agreement at interannual timescales for the Mississippi basin where the hydrological signals are also large.
GPS uplift or subsidence trends which may be caused, next to mass loading, by a plethora of other geophysical or anthropogenic effects are difficult to compare. At the annual timescale no clear picture emerges: gravity solutions from the Swarm reconstruction approach seem to outperform monthly Swarm-only solution for some regions with large signal while for others the Swarm-only monthly solutions appear closer to GPS. Another way of validating the global gravity solutions is by assessing how well they would allow the prediction of the orbits of other satellites.
Lageos data is processed in orbit arcs of 10 days, while for Ajisai, Starlette and Stella 3-day arcs are chosen due to limitations in modeling the atmospheric drag Figure 5 shows the residuals computed with a monthly Swarm-only gravity fields and b the reconstruction approach. As already mentioned, the quality of the monthly Swarm-only solutions is affected by ionospheric disturbances in and 37 , which can clearly be seen in Fig.
SLR observations form Ajisai, Stella and Starlette do not match the Swarm-only gravity field during the beginning of the Swarm mission, leading to residuals of several decimeters. The SLR residuals reveal again that the reconstruction approach does not seem to be affected by the ionospheric disturbances. Starting in , the residuals of both approaches get more similar, but those of the reconstruction are still lower.
Thus, the improvement when using the Swarm reconstruction can mainly been seen when looking at Ajisai, Stella and Starlette. In summary, the analysis of SLR range residuals confirms the capability of the reconstruction approach to improve time-variable gravity fields from Swarm.
Post-fit range residuals of five SLR satellites. The two variants of the new reconstruction approach reduce the variance further and improve Swarm results, as we show during the overlap period with GRACE.
Our main result is that for the recent period mid til mid which is not covered by GRACE, mass change in all major basins occurs quite regularly compared to earlier GRACE results, as could have been predicted from a climatology. The reconstructed maps have a lower RMSE of 0. Onboard GPS instruments determine the exact position of the satellites over the Earth. GRACE measures changes in Earth's gravity field, which are directly related to changes in surface mass.
The surface mass signal largely reflects total water storage TWS ; over the ocean TWS is interpreted as ocean bottom pressure and on land it is the the sum of groundwater, soil moisture, surface water, snow and ice. Large earthquakes can also lead to mass changes and anomalies in the gravity measurements.
The following was contributed by Sean C. Strengths : GRACE measures changes in the Earth's gravity field, which at monthly timescales, are related to changes in surface mass. Whereas some radiometric remote sensing products e. Because GRACE senses changes in surface mass, it can be used to infer changes in land ice, land water storage, and ocean bottom pressure. The raw gravity field data are typically in the form of spherical harmonic coefficients, which contain measurement errors that are non-uniform.
To obtain a useful combination of resolution and accuracy, the data must be filtered or regularized in a least squares sense, which is a form of filtering. The number of filtering strategies, and the lack of information on the filtering method's effect on the signal, can be confusing.
GRACE data have been used to estimate changes in land ice e. These data are being included in land data assimilation schemes and drought monitoring products. GRACE has been used to estimate global ocean mass changes. One of the most significant findings was the extent to which the melting of Earth's polar ice caps has contributed to rising sea levels , Watkins said.
As that ice melts — whether the water becomes a part of the ocean or seeps into the soil — it changes how mass is distributed on our planet. And when Earth's mass distribution changes, so does its gravity field. By measuring those changes, the GRACE missions can track Earth's water cycle and how climate change affects oceans, glaciers, sea ice, groundwater and even moisture in the atmosphere.
0コメント