Atmospheric Correction for [[Interferometric synthetic-aperture radar|Interferometric synthetic aperture radar]] (InSAR) technique is a set of different methods to remove artifact displacement from an Interferogram caused by the effect of weather variables such as [[humidity]], [[temperature]], and [[pressure]].<ref name=":0" /> “Corrections” for atmospheric signals are an important stage of InSAR data processing in many study areas to mesure surface displacment because the displacement result of [[Interferometric synthetic-aperture radar|InSAR]] can become inaccurate by 10–14 cm if relative humidity differs by 20%.<ref name=":0" /> Overall, atmospheric correction methods can be divided into two categories: a) Using Atmospheric Phase Screen (APS) [[Statistics|statistical]] properties and b) Using auxiliary (external) data such as [[Global Positioning System|GPS]] measurement, [[Multispectral imaging|multi-spectral]] observations, local meteorological models, and Global Atmospheric Models.<ref name="ReferenceA">{{Cite journal |last1=Hu |first1=Zhongbo |last2=Mallorquí |first2=Jordi J. |date=January 2019 |title=An Accurate Method to Correct Atmospheric Phase Delay for InSAR with the ERA5 Global Atmospheric Model |journal=Remote Sensing |language=en |volume=11 |issue=17 |pages=1969 |doi=10.3390/rs11171969 |bibcode=2019RemS...11.1969H |issn=2072-4292|doi-access=free }}</ref>
Atmospheric Correction for [[Interferometric synthetic-aperture radar|Interferometric synthetic aperture radar]] ([[Interferometric synthetic-aperture radar|InSAR]]) technique is a set of different methods to remove artifact displacement from an [[Interferometry|Interferogram]] caused by the effect of weather variables such as [[humidity]], [[temperature]], and [[pressure]].<ref name=":0" /> An [[Interferometry|interferogram]] is generated by processing two SAR images before and after a geophysical event like an [[earthquake]]. [[Interferometry|Interferograms]] are similar to a rainbow, and each colour cycle shows a specific displacement based on radar wavelength.<ref>{{Cite book |url=https://link.springer.com/book/10.1007/0-306-47633-9 |title=Radar Interferometry |language=en |doi=10.1007/0-306-47633-9}}</ref>“Corrections” for atmospheric signals are an important stage of InSAR data processing in many study areas to measure surface displacment because the displacement result of [[Interferometric synthetic-aperture radar|InSAR]] can become inaccurate by 10–14 cm if relative humidity differs by 20%.<ref name=":0" /> Overall, atmospheric correction methods can be divided into two categories: a) Using Atmospheric Phase Screen (APS) [[Statistics|statistical]] properties and b) Using auxiliary (external) data such as [[Global Positioning System|GPS]] measurement, [[Multispectral imaging|multi-spectral]] observations, local meteorological models, and Global Atmospheric Models.<ref name="ReferenceA">{{Cite journal |last1=Hu |first1=Zhongbo |last2=Mallorquí |first2=Jordi J. |date=January 2019 |title=An Accurate Method to Correct Atmospheric Phase Delay for InSAR with the ERA5 Global Atmospheric Model |journal=Remote Sensing |language=en |volume=11 |issue=17 |pages=1969 |doi=10.3390/rs11171969 |bibcode=2019RemS...11.1969H |issn=2072-4292|doi-access=free }}</ref>
Interferometric synthetic aperture radar ([[Interferometric synthetic-aperture radar|InSAR]]) can provide accurate (millimeter-level) ground displacement fields for large areas over hundreds of kilometers.<ref name=":0" /> This technique uses two SAR images of the same area acquired at different times to measure surface motion.<ref name=":0" /> Nevertheless, the result of [[Interferometric synthetic-aperture radar|InSAR]] in interferogram form includes actual displacement and other effects. Hence, these other effects must be calculated and removed from interferograms to achieve an accurate result of real ground displacement.<ref>{{Cite book |last1=Xing |first1=Mengdao |url=https://www.mdpi.com/books/book/2809 |title=InSAR Signal and Data Processing |last2=Lu |first2=Zhong |last3=Yu |first3=Hanwen |date=2020-09-04 |publisher=MDPI, Basel |isbn=978-3-03936-984-3 |language=en}}</ref> Some of these errors have more influence, such as orbital errors, topographic effects and atmospheric artefacts. In addition, other noises like unwrapping and geometry errors should be considered to reach a trustable result.<ref name=":0" /> There are several methods to remove all these noises until a reasonable and acceptable amount except atmospheric effects with high accuracy.Thus, a significant challenge for the [[Interferometric synthetic-aperture radar|InSAR]] community remains atmospheric artifacts. Sometimes especially in areas with high humidity, the effect of atmospheric noise is much larger than geophysical events and prevents surface displacement detection.<ref name=":0">{{Cite book |url=https://link.springer.com/book/10.1007/0-306-47633-9 |title=Radar Interferometry |series=Remote Sensing and Digital Image Processing |year=2001 |volume=2 |language=en |doi=10.1007/0-306-47633-9|isbn=978-0-7923-6945-5 }}</ref>
[[Interferometric synthetic-aperture radar|Interferometric synthetic aperture radar]] ([[Interferometric synthetic-aperture radar|InSAR]]) can provide accurate (millimeter-level) ground displacement fields for large areas over hundreds of kilometers.<ref name=":0" /> This technique uses two SAR images of the same area acquired at different times to measure surface motion.<ref name=":0" /> Nevertheless, the result of [[Interferometric synthetic-aperture radar|InSAR]] in [[Interferometry|interferogram]] form includes actual displacement and other effects. Hence, these other effects must be calculated and removed from [[Interferometry|interferograms]] to achieve an accurate result of real ground displacement.<ref>{{Cite book |last1=Xing |first1=Mengdao |url=https://www.mdpi.com/books/book/2809 |title=InSAR Signal and Data Processing |last2=Lu |first2=Zhong |last3=Yu |first3=Hanwen |date=2020-09-04 |publisher=MDPI, Basel |isbn=978-3-03936-984-3 |language=en}}</ref> Some of these errors have more influence, such as orbital errors, topographic effects and atmospheric artefacts. In addition, other noises like unwrapping and geometry errors should be considered to reach a trustable result.<ref name=":0" /> There are several methods to remove all these noises until a reasonable and acceptable amount except atmospheric effects with high accuracy.Thus, a significant challenge for the [[Interferometric synthetic-aperture radar|InSAR]] community remains atmospheric artefacts. Sometimes especially in areas with high humidity, the effect of atmospheric noise is much larger than geophysical events and prevents surface displacement detection.<ref name=":0">{{Cite book |url=https://link.springer.com/book/10.1007/0-306-47633-9 |title=Radar Interferometry |series=Remote Sensing and Digital Image Processing |year=2001 |volume=2 |language=en |doi=10.1007/0-306-47633-9|isbn=978-0-7923-6945-5 }}</ref>
On the other side, atmospheric noise might have some value for atmospheric research in [[Meteorology]] because atmospheric artefact signals are related to water vapour in the troposphere.<ref name=":8" /> The spatial resolution of the InSAR map for C-band satellites like Sentinel-1 without multi-looking is around 20 meters. That means InSAR can measure Precipitable Water Vapor (PWV) in the atmosphere in a 20m grid over hundreds of kilometres, which is much denser than other methods such as GNSS and space-borne passive sensors.<ref name=":1" /> However, the long revisit time of Sentinel-1 (temporal resolution, 12 days) at the moment is the main disadvantage of this technique from the meteorologists' side. Nevertheless, using the capability of InSAR to measure PWV in high spatial resolution is interesting for meteorological research.<ref>{{Cite journal |last1=Mateus |first1=Pedro |last2=Catalao |first2=Joao |last3=Nico |first3=Giovanni |date=May 2017 |title=Sentinel-1 Interferometric SAR Mapping of Precipitable Water Vapor Over a Country-Spanning Area |url=https://ieeexplore.ieee.org/document/7855708 |journal=IEEE Transactions on Geoscience and Remote Sensing |volume=55 |issue=5 |pages=2993–2999 |doi=10.1109/TGRS.2017.2658342 |bibcode=2017ITGRS..55.2993M |s2cid=38018728 |issn=0196-2892}}</ref>
On the other side, [[Atmosphere|atmospheric]] noise might have some value for atmospheric research in [[Meteorology]] because atmospheric artefact signals are related to water vapour in the troposphere.<ref name=":8" /> The spatial resolution of the [[Interferometric synthetic-aperture radar|InSAR]] map for C-band satellites like Sentinel-1 without multi-looking is around 20 meters. That means InSAR can measure Precipitable Water Vapor (PWV) in the [[atmosphere]] in a 20m grid over hundreds of kilometres, which is much denser than other methods such as GNSS and space-borne passive sensors.<ref name=":1" /> However, the long revisit time of Sentinel-1 (temporal resolution, 12 days) at the moment is the main disadvantage of this technique from the meteorologists' side. Nevertheless, using the capability of InSAR to measure PWV in high spatial resolution is interesting for meteorological research.<ref>{{Cite journal |last1=Mateus |first1=Pedro |last2=Catalao |first2=Joao |last3=Nico |first3=Giovanni |date=May 2017 |title=Sentinel-1 Interferometric SAR Mapping of Precipitable Water Vapor Over a Country-Spanning Area |url=https://ieeexplore.ieee.org/document/7855708 |journal=IEEE Transactions on Geoscience and Remote Sensing |volume=55 |issue=5 |pages=2993–2999 |doi=10.1109/TGRS.2017.2658342 |bibcode=2017ITGRS..55.2993M |s2cid=38018728 |issn=0196-2892}}</ref>
== Atmospheric noise ==
== Atmospheric noise ==
Line 12:
Line 12:
=== Ionosphere ===
=== Ionosphere ===
Ionospheric phase noise, which occurs more apparent with larger wavelengths, such as P or L-band radars, is a consequence of variations in free electron density in the 100–1000 km altitude along the travel path.<ref name=":9">{{Cite journal |last1=Gray |first1=A. Laurence |last2=Mattar |first2=Karim E. |last3=Sofko |first3=George |date=2000-05-15 |title=Influence of ionospheric electron density fluctuations on satellite radar interferometry |url=http://doi.wiley.com/10.1029/2000GL000016 |journal=Geophysical Research Letters |language=en |volume=27 |issue=10 |pages=1451–1454 |bibcode=2000GeoRL..27.1451G |doi=10.1029/2000GL000016 |s2cid=130384140}}</ref> Radar satellites with large wavelengths like ALOS 1/2 Polsar (L-band, λ=~24 cm) and [[NISAR (satellite)|NISAR]] are more vulnerable to ionospheric delays. However, this noise is roughly less effective in C-Band (λ=5.6 cm) and X-band (λ=3.1 cm) SAR satellites such as Envisat, RADARSAT, ERS1/2 and TerraSAR-X and usually negligible.<ref name=":9" /> There are a couple of methods to remove the Ionospheric noise artifact:
[[Ionosphere|Ionospheric]] phase noise, which occurs more apparent with larger [[Wavelength|wavelengths]], such as P or L-band radars, is a consequence of variations in free electron density in the 100–1000 km altitude along the travel path.<ref name=":9">{{Cite journal |last1=Gray |first1=A. Laurence |last2=Mattar |first2=Karim E. |last3=Sofko |first3=George |date=2000-05-15 |title=Influence of ionospheric electron density fluctuations on satellite radar interferometry |url=http://doi.wiley.com/10.1029/2000GL000016 |journal=Geophysical Research Letters |language=en |volume=27 |issue=10 |pages=1451–1454 |bibcode=2000GeoRL..27.1451G |doi=10.1029/2000GL000016 |s2cid=130384140}}</ref> Radar satellites with large wavelengths like [[Advanced Land Observation Satellite|ALOS Polsar]]1/2 (L-band, λ=~24 cm) and [[NISAR (satellite)|NISAR]] are more vulnerable to [[Ionosphere|ionospheric]] delays. However, this noise is roughly less effective in C-Band (λ=5.6 cm) and X-band (λ=3.1 cm) SAR satellites such as Envisat, [[Radarsat-1|RADARSAT]], ERS1/2 and TerraSAR-X and usually negligible.<ref name=":9" /> There are a couple of methods to remove the Ionospheric noise artifact:
# Using auxiliary data like [[Satellite navigation|GNSS]]<ref>{{Cite journal |last1=Chen |first1=Jingyi |last2=Zebker |first2=Howard A. |date=April 2012 |title=Ionospheric Artifacts in Simultaneous L-Band InSAR and GPS Observations |url=https://ieeexplore.ieee.org/document/6035969 |journal=IEEE Transactions on Geoscience and Remote Sensing |volume=50 |issue=4 |pages=1227–1239 |doi=10.1109/TGRS.2011.2164805 |bibcode=2012ITGRS..50.1227C |s2cid=5769312 |issn=1558-0644}}</ref>
# Using auxiliary data like [[Satellite navigation|GNSS]]<ref>{{Cite journal |last1=Chen |first1=Jingyi |last2=Zebker |first2=Howard A. |date=April 2012 |title=Ionospheric Artifacts in Simultaneous L-Band InSAR and GPS Observations |url=https://ieeexplore.ieee.org/document/6035969 |journal=IEEE Transactions on Geoscience and Remote Sensing |volume=50 |issue=4 |pages=1227–1239 |doi=10.1109/TGRS.2011.2164805 |bibcode=2012ITGRS..50.1227C |s2cid=5769312 |issn=1558-0644}}</ref>
Line 21:
Line 21:
=== Troposphere ===
=== Troposphere ===
The troposphere is the lower layer of the atmosphere, with up to 90% of the atmosphere's water vapor. A height of 1.4 km from the ground surface contains 50% of the water vapor mass, and both on the global and local scales, most of the world's weather takes place in this layer.<ref name=":1">{{Cite journal |last1=Mateus |first1=Pedro |last2=Nico |first2=Giovanni |last3=Catalao |first3=Joao |date=December 2015 |title=Uncertainty Assessment of the Estimated Atmospheric Delay Obtained by a Numerical Weather Model (NMW) |url=https://ieeexplore.ieee.org/document/7150541 |journal=IEEE Transactions on Geoscience and Remote Sensing |volume=53 |issue=12 |pages=6710–6717 |doi=10.1109/TGRS.2015.2446758 |bibcode=2015ITGRS..53.6710M |s2cid=8147781 |issn=0196-2892}}</ref> Therefore, the tropospheric path delays are caused by differences in temperature, atmospheric pressure, and the water vapor in the lower part of the atmosphere. Microwave atmospheric delay is a sum of the turbulent and vertical stratified components. The turbulent component can affect flat and mountainous terrains and is usually highly correlated in space and uncorrelated in time. On the other hand, the vertical stratified component with a high correlation to topography only influences areas with topography variations, such as hills or mountains.<ref name=":0" /><ref>{{Cite journal |last1=Xiao |first1=Ruya |last2=Yu |first2=Chen |last3=Li |first3=Zhenhong |last4=He |first4=Xiufeng |date=April 2021 |title=Statistical assessment metrics for InSAR atmospheric correction: Applications to generic atmospheric correction online service for InSAR (GACOS) in Eastern China |url=https://linkinghub.elsevier.com/retrieve/pii/S0303243420309326 |journal=International Journal of Applied Earth Observation and Geoinformation |language=en |volume=96 |pages=102289 |doi=10.1016/j.jag.2020.102289|bibcode=2021IJAEO..9602289X |s2cid=231885794 }}</ref>
The [[troposphere]] is the lower layer of the atmosphere, with up to 90% of the atmosphere's water vapor. A height of 1.4 km from the ground surface contains 50% of the water vapor mass, and both on the global and local scales, most of the world's weather takes place in this layer.<ref name=":1">{{Cite journal |last1=Mateus |first1=Pedro |last2=Nico |first2=Giovanni |last3=Catalao |first3=Joao |date=December 2015 |title=Uncertainty Assessment of the Estimated Atmospheric Delay Obtained by a Numerical Weather Model (NMW) |url=https://ieeexplore.ieee.org/document/7150541 |journal=IEEE Transactions on Geoscience and Remote Sensing |volume=53 |issue=12 |pages=6710–6717 |doi=10.1109/TGRS.2015.2446758 |bibcode=2015ITGRS..53.6710M |s2cid=8147781 |issn=0196-2892}}</ref> Therefore, the tropospheric path delays are caused by differences in [[temperature]], [[Atmosphere|atmospheric]] [[pressure]], and the water vapor in the lower part of the atmosphere. Microwave atmospheric delay is a sum of the turbulent and vertical stratified components. The turbulent component can affect flat and mountainous terrains and is usually highly correlated in space and uncorrelated in time. On the other hand, the vertical stratified component with a high correlation to [[topography]] only influences areas with topography variations, such as hills or mountains.<ref name=":0" /><ref>{{Cite journal |last1=Xiao |first1=Ruya |last2=Yu |first2=Chen |last3=Li |first3=Zhenhong |last4=He |first4=Xiufeng |date=April 2021 |title=Statistical assessment metrics for InSAR atmospheric correction: Applications to generic atmospheric correction online service for InSAR (GACOS) in Eastern China |url=https://linkinghub.elsevier.com/retrieve/pii/S0303243420309326 |journal=International Journal of Applied Earth Observation and Geoinformation |language=en |volume=96 |pages=102289 |doi=10.1016/j.jag.2020.102289|bibcode=2021IJAEO..9602289X |s2cid=231885794 }}</ref>
Although the tropospheric effect is regarded as noise in the InSAR community, it has great advantages in meteorology and enables scientists to predict and model water vapor in the troposphere.<ref name=":8" />
Although the tropospheric effect is regarded as noise in the [[Interferometric synthetic-aperture radar|InSAR]] community, it has great advantages in [[meteorology]] and enables scientists to predict and model water vapor in the [[troposphere]].<ref name=":8" />
The propagationtion delay of electromagnetic waves of radar satellites in the troposphere will be explained in the following in terms of meteorology and the InSAR side.<ref name=":0" />
The propagationtion delay of electromagnetic waves of radar satellites in the [[troposphere]] will be explained in the following in terms of [[meteorology]] and the InSAR side.<ref name=":0" />
[[File:Ss5456sssss.jpg|alt=Layers of troposphere and amount of water vapor in each altitude|center|thumb|518x518px|''Amount of water vapor in each altitude of the troposphere. Around 90% of water vapor is contained in the troposphere and 50% of which is between the earth's surface up to about a height of 1.4 km'']]
[[File:Ss5456sssss.jpg|alt=Layers of troposphere and amount of water vapor in each altitude|center|thumb|518x518px|''Amount of water vapor in each altitude of the troposphere. Around 90% of water vapor is contained in the troposphere and 50% of which is between the earth's surface up to about a height of 1.4 km'']]
==== Troposphere effect on InSAR ====
==== Troposphere effect on [[Interferometric synthetic-aperture radar|InSAR]] ====
From the InSAR side, signal delay caused by variations of tropospheric properties in space and time is the source of major challenges for InSAR rechnique. In other words, tropospheric perturbation, caused by the differences in relative humidity, temperature, and pressure in the lower part of the troposphere between two acquisitions, can lead to additional noise in form of fringes up to 15–20 cm on interferograms.<ref name=":2">{{Cite journal |last1=Bekaert |first1=D.P.S. |last2=Walters |first2=R.J. |last3=Wright |first3=T.J. |last4=Hooper |first4=A.J. |last5=Parker |first5=D.J. |date=December 2015 |title=Statistical comparison of InSAR tropospheric correction techniques |url=https://linkinghub.elsevier.com/retrieve/pii/S0034425715301231 |journal=Remote Sensing of Environment |language=en |volume=170 |pages=40–47 |doi=10.1016/j.rse.2015.08.035|bibcode=2015RSEnv.170...40B }}</ref> The atmospheric noise on InSAR results can include a wide range of wavelengths (short to long). Long wavelength errors, usually seen as a ramp (like a trend) in interferograms, are caused by changing the weather system in the study area very smoothly between two SAR images. Since this noise is similar to orbital ramp error and solid earth tides, detecting that in an interferogram is complicated.<ref>{{Cite journal |last1=Shirzaei |first1=Manoochehr |last2=Walter |first2=Thomas R. |date=November 2011 |title=Estimating the Effect of Satellite Orbital Error Using Wavelet-Based Robust Regression Applied to InSAR Deformation Data |url=https://ieeexplore.ieee.org/document/5779729 |journal=IEEE Transactions on Geoscience and Remote Sensing |volume=49 |issue=11 |pages=4600–4605 |doi=10.1109/TGRS.2011.2143419 |bibcode=2011ITGRS..49.4600S |s2cid=19391765 |issn=1558-0644}}</ref> On the other hand, rapidly changing weather in a small area can cause artefact signals that correlate with topography because water vapour variation in surface and altitudes is different. Moreover, a rain cloud in a small region can generate a turbulence error which would be visible like uplift or subsidence on interferogram.<ref name=":3" /> Overall, the tropospheric error on interferogram can be classified into space and time:
From the [[Interferometric synthetic-aperture radar|InSAR]] side, signal delay caused by variations of [[Troposphere|tropospheric]] properties in space and time is the source of major challenges for [[Interferometric synthetic-aperture radar|InSAR]] technique. In other words, tropospheric perturbation, caused by the differences in relative [[humidity]], [[temperature]], and [[pressure]] in the lower part of the [[troposphere]] between two acquisitions, can lead to additional noise in form of fringes up to 15–20 cm on [[Interferometry|interferograms]].<ref name=":2">{{Cite journal |last1=Bekaert |first1=D.P.S. |last2=Walters |first2=R.J. |last3=Wright |first3=T.J. |last4=Hooper |first4=A.J. |last5=Parker |first5=D.J. |date=December 2015 |title=Statistical comparison of InSAR tropospheric correction techniques |url=https://linkinghub.elsevier.com/retrieve/pii/S0034425715301231 |journal=Remote Sensing of Environment |language=en |volume=170 |pages=40–47 |doi=10.1016/j.rse.2015.08.035|bibcode=2015RSEnv.170...40B }}</ref> The atmospheric noise on [[Interferometric synthetic-aperture radar|InSAR]] results can include a wide range of wavelengths (short to long). Long wavelength errors, usually seen as a ramp (like a trend) in interferograms, are caused by changing the weather system in the study area very smoothly between two SAR images. Since this noise is similar to orbital ramp error and solid earth tides, detecting that in an [[Interferometry|interferogram]] is complicated.<ref>{{Cite journal |last1=Shirzaei |first1=Manoochehr |last2=Walter |first2=Thomas R. |date=November 2011 |title=Estimating the Effect of Satellite Orbital Error Using Wavelet-Based Robust Regression Applied to InSAR Deformation Data |url=https://ieeexplore.ieee.org/document/5779729 |journal=IEEE Transactions on Geoscience and Remote Sensing |volume=49 |issue=11 |pages=4600–4605 |doi=10.1109/TGRS.2011.2143419 |bibcode=2011ITGRS..49.4600S |s2cid=19391765 |issn=1558-0644}}</ref> On the other hand, rapidly changing weather in a small area can cause artefact signals that correlate with topography because water vapour variation in surface and altitudes is different. Moreover, a rain cloud in a small region can generate a turbulence error which would be visible like uplift or subsidence on [[Interferometry|interferogram]].<ref name=":3" /> Overall, the [[Troposphere|tropospheric]] error on [[Interferometry|interferogram]] can be classified into space and time<ref name=":3" />:
{|
{|
! rowspan="4" |Tropospheric artifact
! rowspan="4" |Tropospheric artifact
Line 46:
Line 46:
|}
|}
[[File:Reflection0010013s.tif|thumb|395x395px|These figures show the correlation between the signal delay and altitude in two points( p and q ) with different hight at t1(date of image one) and t2(date of image two). The signal reflected from q would have more delay because the amount of water vapor near the surface is very high and subject to more variation in t2.<ref name=":10">{{Cite journal |last1=Havazli |first1=Emre |last2=Wdowinski |first2=Shimon |date=January 2021 |title=Detection Threshold Estimates for InSAR Time Series: A Simulation of Tropospheric Delay Approach |journal=Sensors |language=en |volume=21 |issue=4 |pages=1124 |doi=10.3390/s21041124 |pmid=33562774 |pmc=7915624 |bibcode=2021Senso..21.1124H |issn=1424-8220|doi-access=free }}</ref>]]
[[File:Reflection0010013s.tif|thumb|395x395px|These figures show the correlation between the signal delay and altitude in two points( p and q ) with different hight at t1(date of image one) and t2(date of image two). The signal reflected from q would have more delay because the amount of water vapor near the surface is very high and subject to more variation in t2.<ref name=":10">{{Cite journal |last1=Havazli |first1=Emre |last2=Wdowinski |first2=Shimon |date=January 2021 |title=Detection Threshold Estimates for InSAR Time Series: A Simulation of Tropospheric Delay Approach |journal=Sensors |language=en |volume=21 |issue=4 |pages=1124 |doi=10.3390/s21041124 |pmid=33562774 |pmc=7915624 |bibcode=2021Senso..21.1124H |issn=1424-8220|doi-access=free }}</ref>]]
* '''''Vertical stratification:''''' Vertical stratification is correlated with topography and can be seen wherever there are hills or mountains because the amount of water vapour is not constant near the surface and changes significantly with altitude. Hence, changing weather (water vapour) between two SAR images can cause delay and, thus, vertical stratification noise in the interferogram.<ref name=":10" />
* '''''Vertical stratification:''''' Vertical stratification is correlated with [[topography]] and can be seen wherever there are hills or mountains because the amount of water vapour is not constant near the surface and changes significantly with altitude. Hence, changing weather (water vapour) between two SAR images can cause delay and, thus, vertical stratification noise in the interferogram.<ref name=":10" />
* '''''Turbulence:''''' Turbulent mixing consist of a short-scale (few km) component and is not directly correlated with [[topography]] and affects both flat terrains and higher altitude terrains.<ref name=":10" />
* '''''Turbulence:''''' Turbulent mixing consists of a short-scale (few km) components and is not directly correlated with [[topography]] and affects both flat terrains and higher altitude terrains.<ref name=":10" />
*
*
*[[File:Teneaw.jpg|thumb|564x564px|Popocatepetel, Mexico, 2014-2017 summit displacement and seasonal effect during 2016 (Adopted from Falk Amelung presentation, InSAR Meteorology Miami 2018<ref name=":8">{{Cite web |title=InSARMeteorologyMiami2018 |url=https://insarmeteorologymiami2018.wordpress.com/ |access-date=2022-11-06 |website=InSARMeteorologyMiami2018 |language=en}}</ref>).]]'''''Seasonal (systematic) effect''''' : Variation of moisture content of the atmosphere during different seasons may impose an artifact and bias the InSAR displacement time series. The large number of SAR acquisitions will not reduce the velocity bias necessarily.<ref name=":3">{{Cite journal |last1=Fattahi |first1=Heresh |last2=Amelung |first2=Falk |date=December 2015 |title=InSAR bias and uncertainty due to the systematic and stochastic tropospheric delay |url=https://onlinelibrary.wiley.com/doi/abs/10.1002/2015JB012419 |journal=Journal of Geophysical Research: Solid Earth |language=en |volume=120 |issue=12 |pages=8758–8773 |doi=10.1002/2015JB012419 |bibcode=2015JGRB..120.8758F |s2cid=19372953 |issn=2169-9313}}</ref> Nonetheless, the stratified tropospheric delay correction can mitigate the effects of the seasonal delay.<ref name=":3" /><ref name=":4">{{Cite journal |last1=Jolivet |first1=R. |last2=Grandin |first2=R. |last3=Lasserre |first3=C. |last4=Doin |first4=M.-P. |last5=Peltzer |first5=G. |date=September 2011 |title=Systematic InSAR tropospheric phase delay corrections from global meteorological reanalysis data: CORRECTING INSAR WITH ERA-INTERIM |url=http://doi.wiley.com/10.1029/2011GL048757 |journal=Geophysical Research Letters |language=en |volume=38 |issue=17 |pages=n/a |doi=10.1029/2011GL048757|s2cid=12904863 }}</ref>
*[[File:Teneaw.jpg|thumb|564x564px|Popocatepetel, Mexico, 2014-2017 summit displacement and seasonal effect during 2016 (Adopted from Falk Amelung presentation, InSAR Meteorology Miami 2018<ref name=":8">{{Cite web |title=InSARMeteorologyMiami2018 |url=https://insarmeteorologymiami2018.wordpress.com/ |access-date=2022-11-06 |website=InSARMeteorologyMiami2018 |language=en}}</ref>).]]'''''Seasonal (systematic) effect''''' : Variation of moisture content of the atmosphere during different seasons may impose an artifact and bias the InSAR displacement time series. The large number of SAR acquisitions will not reduce the velocity bias necessarily.<ref name=":3">{{Cite journal |last1=Fattahi |first1=Heresh |last2=Amelung |first2=Falk |date=December 2015 |title=InSAR bias and uncertainty due to the systematic and stochastic tropospheric delay |url=https://onlinelibrary.wiley.com/doi/abs/10.1002/2015JB012419 |journal=Journal of Geophysical Research: Solid Earth |language=en |volume=120 |issue=12 |pages=8758–8773 |doi=10.1002/2015JB012419 |bibcode=2015JGRB..120.8758F |s2cid=19372953 |issn=2169-9313}}</ref> Nonetheless, the stratified tropospheric delay correction can mitigate the effects of the seasonal delay.<ref name=":3" /><ref name=":4">{{Cite journal |last1=Jolivet |first1=R. |last2=Grandin |first2=R. |last3=Lasserre |first3=C. |last4=Doin |first4=M.-P. |last5=Peltzer |first5=G. |date=September 2011 |title=Systematic InSAR tropospheric phase delay corrections from global meteorological reanalysis data: CORRECTING INSAR WITH ERA-INTERIM |url=http://doi.wiley.com/10.1029/2011GL048757 |journal=Geophysical Research Letters |language=en |volume=38 |issue=17 |pages=n/a |doi=10.1029/2011GL048757|s2cid=12904863 }}</ref>
==== Troposphere effect on Meteorology ====
==== Troposphere effect on [[Meteorology]] ====
In terms of Meteorology, tropospheric delay interestingly can be regarded as a useful tool for Meteorology purposes.<ref name=":8" /> Traditional methods for measuring water vapor include using a) [[Meteorology|Meteorological]] stations, b) [[Radiosonde]], c) [[Spectrometer]], and d) [[Global Positioning System|GPS]].In recent years, InSAR has enabled meteorological scientists to measure precipitable water vapour (PWV) in the atmosphere with high spatial resolution.<ref name=":1" /> Although InSAR temporal sampling (12 days for sentinel1) of atmosphere properties is coarser than the GPS (15 min), the main advantage of InSAR PWV maps is its high spatial resolution of up to a few meters (20 m).<ref name=":1" /> Therefore, [[Meteorology|meteorological]] scientists combine traditional [[Global Positioning System|GPS]] tomography of [[atmosphere]] with InSAR data (atmospheric part) to increase their models' temporal and spatial resolution.<ref>{{Cite journal |last1=Benevides |first1=Pedro |last2=Nico |first2=Giovanni |last3=Catalao |first3=Joao |last4=Miranda |first4=P. M. A. |date=February 2016 |title=Bridging InSAR and GPS Tomography: A New Differential Geometrical Constraint |url=https://ieeexplore.ieee.org/document/7293637 |journal=IEEE Transactions on Geoscience and Remote Sensing |volume=54 |issue=2 |pages=697–702 |doi=10.1109/TGRS.2015.2463263 |bibcode=2016ITGRS..54..697B |s2cid=29078915 |issn=0196-2892}}</ref>
In terms of [[Meteorology]], tropospheric delay interestingly can be regarded as a useful tool for Meteorology purposes.<ref name=":8" /> Traditional methods for measuring water vapor include using a) [[Meteorology|Meteorological]] stations, b) [[Radiosonde]], c) [[Spectrometer]], and d) [[Global Positioning System|GPS]].In recent years, InSAR has enabled meteorological scientists to measure precipitable water vapour (PWV) in the atmosphere with high spatial resolution.<ref name=":1" /> Although InSAR temporal sampling (12 days for sentinel1) of atmosphere properties is coarser than the GPS (15 min), the main advantage of InSAR PWV maps is its high spatial resolution of up to a few meters (20 m).<ref name=":1" /> Therefore, [[Meteorology|meteorological]] scientists combine traditional [[Global Positioning System|GPS]] tomography of [[atmosphere]] with InSAR data (atmospheric part) to increase their models' temporal and spatial resolution.<ref>{{Cite journal |last1=Benevides |first1=Pedro |last2=Nico |first2=Giovanni |last3=Catalao |first3=Joao |last4=Miranda |first4=P. M. A. |date=February 2016 |title=Bridging InSAR and GPS Tomography: A New Differential Geometrical Constraint |url=https://ieeexplore.ieee.org/document/7293637 |journal=IEEE Transactions on Geoscience and Remote Sensing |volume=54 |issue=2 |pages=697–702 |doi=10.1109/TGRS.2015.2463263 |bibcode=2016ITGRS..54..697B |s2cid=29078915 |issn=0196-2892}}</ref>
[[File:Ppt file Seminar 28.jpg|alt=|thumb|588x588px|Schematic resolution comparison of InSAR and GPS for a large area. Red points in InSAR and GPS show pixel resolution, and in GPS stations, respectively.|center]]
[[File:Ppt file Seminar 28.jpg|alt=|thumb|588x588px|Schematic resolution comparison of InSAR and GPS for a large area. Red points in InSAR and GPS show pixel resolution, and in GPS stations, respectively.|center]]
== Tropospheric correction methods ==
== Tropospheric correction methods ==
The phase delay through the troposphere can be characterized by the refractivity (N).
The phase delay through the [[troposphere]] can be characterized by the [[Refractive index|refractivity]] (N).
<math>\Delta_{atm}=10^-6\int N(z)dz</math>[[File:Asdssadasdasdasdasdasd.jpg|Schematic representation of the electromagnetic waves delays in tropospheric based on dry and wet components. Zenith Total Delay (ZTD) correspond to total delay, Zenith Wet Delay (ZWD) corresponds to Wet delay and Zenith Hydrostatic Delay (ZHD) is related to Dry delay of troposphere.|alt=Figure 4. Schematic representation of the electromagnetic waves delays in tropospheric based on dry and wet components. Zenith Total Delay (ZTD) correspond to total delay, Zenith Wet Delay (ZWD) corresponds to Wet delay and Zenith Hydrostatic Delay (ZHD) is related to Dry delay of troposphere.|thumb|217x217px]]The atmospheric parameters can be used to calculate N(z) as the atmosphere's refractivity.<ref name=":11"/> This delay can be divided into two components: Dry delay and Wet delay. The dry delay, which is determined by the temperature and pressure of dry air, and the wet delay, which is determined mainly by the amount of water vapor in the troposphere.<ref name=":1" />
<math>\Delta_{atm}=10^-6\int N(z)dz</math>[[File:Asdssadasdasdasdasdasd.jpg|Schematic representation of the electromagnetic waves delays in tropospheric based on dry and wet components. Zenith Total Delay (ZTD) correspond to total delay, Zenith Wet Delay (ZWD) corresponds to Wet delay and Zenith Hydrostatic Delay (ZHD) is related to Dry delay of troposphere.|alt=Figure 4. Schematic representation of the electromagnetic waves delays in tropospheric based on dry and wet components. Zenith Total Delay (ZTD) correspond to total delay, Zenith Wet Delay (ZWD) corresponds to Wet delay and Zenith Hydrostatic Delay (ZHD) is related to Dry delay of troposphere.|thumb|217x217px]]The atmospheric parameters can be used to calculate N(z) as the atmosphere's [[Refractive index|refractivity]].<ref name=":11"/> This delay can be divided into two components: Dry delay and Wet delay. The dry delay, which is determined by the temperature and pressure of dry air, and the wet delay, which is determined mainly by the amount of water vapor in the [[troposphere]].<ref name=":1" />
Since there is no accurate technique that could determine the total refractivity on the same spatial scale and temporal sampling as the interferogram itself, various methods have been developed to mesure for the atmospheric contribution within interferograms. In the following a couple of methods are introduced to mitigate tropospheric artifact from interfrograms:
Since there is no accurate technique that could determine the total refractivity on the same spatial scale and temporal sampling as the interferogram itself, various methods have been developed to measure for the atmospheric contribution within interferograms. In the following a couple of methods are introduced to mitigate tropospheric artifacts from interferograms:
=== <small>Time series methods such as Persistent Scatter (PS) and Small Baseline Subset (SBAS)</small> ===
=== <small>Time series methods such as Persistent Scatter (PS) and Small Baseline Subset (SBAS)</small> ===
Stacking has been one of the helpful strategies for reducing the tropospheric artefacts from interferograms.<ref name=":3" /> The stacking method processes several to hundreds of filtered and unwrapped interferograms together using different algorithms like Small Baseline Subset (SBAS) and persistent scatterer (PS) to measure surface displacement during a long period.<ref>{{Cite journal |last=Hooper |first=Andrew |date=2008-08-19 |title=A multi-temporal InSAR method incorporating both persistent scatterer and small baseline approaches |url=http://doi.wiley.com/10.1029/2008GL034654 |journal=Geophysical Research Letters |language=en |volume=35 |issue=16 |pages=L16302 |doi=10.1029/2008GL034654 |bibcode=2008GeoRL..3516302H |s2cid=12817093 |issn=0094-8276}}</ref><ref>{{Cite journal |last1=Tizzani |first1=P. |last2=Berardino |first2=P. |last3=Casu |first3=F. |last4=Euillades |first4=P. |last5=Manzo |first5=M. |last6=Ricciardi |first6=G. P. |last7=Zeni |first7=G. |last8=Lanari |first8=R. |date=2007-06-15 |title=Surface deformation of Long Valley caldera and Mono Basin, California, investigated with the SBAS-InSAR approach |url=https://www.sciencedirect.com/science/article/pii/S0034425706004688 |journal=Remote Sensing of Environment |language=en |volume=108 |issue=3 |pages=277–289 |doi=10.1016/j.rse.2006.11.015 |bibcode=2007RSEnv.108..277T |issn=0034-4257}}</ref> These methods increase the signal-to-noise ratio of signals without any external information and are very useful for mitigating turbulence errors.<ref name=":3" /><ref>{{Cite journal |last1=Hooper |first1=Andrew |last2=Zebker |first2=Howard |last3=Segall |first3=Paul |last4=Kampes |first4=Bert |date=2004-12-16 |title=A new method for measuring deformation on volcanoes and other natural terrains using InSAR persistent scatterers: A NEW PERSISTENT SCATTERERS METHOD |url=http://doi.wiley.com/10.1029/2004GL021737 |journal=Geophysical Research Letters |language=en |volume=31 |issue=23 |doi=10.1029/2004GL021737|s2cid=55418041 }}</ref><ref>{{Cite journal |last1=Berardino |first1=P. |last2=Fornaro |first2=G. |last3=Lanari |first3=R. |last4=Sansosti |first4=E. |date=November 2002 |title=A new algorithm for surface deformation monitoring based on small baseline differential SAR interferograms |url=https://ieeexplore.ieee.org/document/1166596 |journal=IEEE Transactions on Geoscience and Remote Sensing |language=en |volume=40 |issue=11 |pages=2375–2383 |doi=10.1109/TGRS.2002.803792 |bibcode=2002ITGRS..40.2375B |s2cid=11802957 |issn=0196-2892}}</ref> The other advantage of this method is independent on external data and straightforward to implement.
Stacking has been one of the helpful strategies for reducing the tropospheric artefacts from interferograms.<ref name=":3" /> The stacking method processes several to hundreds of filtered and unwrapped interferograms together using different algorithms like Small Baseline Subset (SBAS) and persistent scatterer (PS) to measure surface displacement during a long period.<ref>{{Cite journal |last=Hooper |first=Andrew |date=2008-08-19 |title=A multi-temporal InSAR method incorporating both persistent scatterer and small baseline approaches |url=http://doi.wiley.com/10.1029/2008GL034654 |journal=Geophysical Research Letters |language=en |volume=35 |issue=16 |pages=L16302 |doi=10.1029/2008GL034654 |bibcode=2008GeoRL..3516302H |s2cid=12817093 |issn=0094-8276}}</ref><ref>{{Cite journal |last1=Tizzani |first1=P. |last2=Berardino |first2=P. |last3=Casu |first3=F. |last4=Euillades |first4=P. |last5=Manzo |first5=M. |last6=Ricciardi |first6=G. P. |last7=Zeni |first7=G. |last8=Lanari |first8=R. |date=2007-06-15 |title=Surface deformation of Long Valley caldera and Mono Basin, California, investigated with the SBAS-InSAR approach |url=https://www.sciencedirect.com/science/article/pii/S0034425706004688 |journal=Remote Sensing of Environment |language=en |volume=108 |issue=3 |pages=277–289 |doi=10.1016/j.rse.2006.11.015 |bibcode=2007RSEnv.108..277T |issn=0034-4257}}</ref> These methods increase the signal-to-noise ratio of signals without any external information and are very useful for mitigating turbulence errors.<ref name=":3" /><ref>{{Cite journal |last1=Hooper |first1=Andrew |last2=Zebker |first2=Howard |last3=Segall |first3=Paul |last4=Kampes |first4=Bert |date=2004-12-16 |title=A new method for measuring deformation on volcanoes and other natural terrains using InSAR persistent scatterers: A NEW PERSISTENT SCATTERERS METHOD |url=http://doi.wiley.com/10.1029/2004GL021737 |journal=Geophysical Research Letters |language=en |volume=31 |issue=23 |doi=10.1029/2004GL021737|s2cid=55418041 }}</ref><ref>{{Cite journal |last1=Berardino |first1=P. |last2=Fornaro |first2=G. |last3=Lanari |first3=R. |last4=Sansosti |first4=E. |date=November 2002 |title=A new algorithm for surface deformation monitoring based on small baseline differential SAR interferograms |url=https://ieeexplore.ieee.org/document/1166596 |journal=IEEE Transactions on Geoscience and Remote Sensing |language=en |volume=40 |issue=11 |pages=2375–2383 |doi=10.1109/TGRS.2002.803792 |bibcode=2002ITGRS..40.2375B |s2cid=11802957 |issn=0196-2892}}</ref> The other advantage of this method is independent on external data and straightforward to implement.
''<u>Limitation</u>'': For time series analysis approaches such as PS and SBAS, a large dataset has to be processed and this method moslty can reduce turbulent part of troposphere noise and signals of interest may be incorrectly removed. Thus, this method cannot be implemented only for one interferogram.<ref name=":3" />
''<u>Limitation</u>'': For time series analysis approaches such as PS and SBAS, a large dataset has to be processed and this method mostly can reduce turbulent part of troposphere noise and signals of interest may be incorrectly removed. Thus, this method cannot be implemented only for one interferogram.<ref name=":3" />
=== <small>Estimated tropospheric delay from GNSS</small> ===
=== <small>Estimated tropospheric delay from GNSS</small> ===
[[Satellite navigation|GNSS]] uses microwave portion of the electromagnetic spectrum similar to SAR technique and both measurements are affected by the atmospheric delay and both provide geodetic measurements with comparable accuracy. Therefore, using interpolating GNSS observation (ِdense GNSS network) to estimate tropospheric delay can be a more accurate strategy to correct InSAR observations.<ref>{{Cite journal |last1=Onn |first1=F. |last2=Zebker |first2=H. A. |date=2006 |title=Correction for interferometric synthetic aperture radar atmospheric phase artifacts using time series of zenith wet delay observations from a GPS network |url=http://doi.wiley.com/10.1029/2005JB004012 |journal=Journal of Geophysical Research |language=en |volume=111 |issue=B9 |pages=B09102 |doi=10.1029/2005JB004012 |bibcode=2006JGRB..111.9102O |issn=0148-0227}}</ref>
[[Satellite navigation|GNSS]] uses microwave portion of the electromagnetic spectrum similar to SAR technique and both measurements are affected by the atmospheric delay and both provide geodetic measurements with comparable accuracy. Therefore, using interpolating GNSS observation (ِdense GNSS network) to estimate tropospheric delay can be a more accurate strategy to correct InSAR observations.<ref>{{Cite journal |last1=Onn |first1=F. |last2=Zebker |first2=H. A. |date=2006 |title=Correction for interferometric synthetic aperture radar atmospheric phase artifacts using time series of zenith wet delay observations from a GPS network |url=http://doi.wiley.com/10.1029/2005JB004012 |journal=Journal of Geophysical Research |language=en |volume=111 |issue=B9 |pages=B09102 |doi=10.1029/2005JB004012 |bibcode=2006JGRB..111.9102O |issn=0148-0227}}</ref>
''<u>Limitation</u>'': Interpolation of zenith delay measurements requires dense GNSS network in the study area but GNSS stations are still sparsely distributed or even absent in many regions as well as this method only samples troposphere in the vicinity of individual GNSS sites. Moreover, GNSS datasets are still not freely available to the public in the many areas in the world.<ref name=":12" />
''<u>Limitation</u>'': Interpolation of zenith delay measurements requires dense GNSS network in the study area but GNSS stations are still sparsely distributed or even absent in many regions as well as this method only samples troposphere in the vicinity of individual GNSS sites. Moreover, [[Satellite navigation|GNSS]] datasets are still not freely available to the public in the many areas in the world.<ref name=":12" />
=== <small>Spectrometers Satellite (MERIS and [[Moderate Resolution Imaging Spectroradiometer|MODIS]])</small> ===
=== <small>Spectrometers Satellite (MERIS and [[Moderate Resolution Imaging Spectroradiometer|MODIS]])</small> ===
Advantage of this method is insensitive to the presence of clouds and Global/regional/ local coverage. By using the ERA, uncertainty in ZWD in high latitudes (> 30∘), reach 1–2 cm, while uncertainty in low latitudes (< 30∘), is about 2–6 cm. Most of studies reported varying degrees of success for using this method.<ref name=":4" /> Nonetheless, ERA-interim, and ERA5 global re-analysis models are popular models provided by the European Center for Medium-Range Weather Forecasting (ECMWF) and the HRES ECMWF forecast model.<ref>{{Cite journal |last1=Eff-Darwich |first1=Antonio |last2=Pérez |first2=Juan C. |last3=Fernández |first3=José |last4=García-Lorenzo |first4=Begoña |last5=González |first5=Albano |last6=González |first6=Pablo J. |date=2012-08-01 |title=Using a Mesoscale Meteorological Model to Reduce the Effect of Tropospheric Water Vapour from DInSAR Data: A Case Study for the Island of Tenerife, Canary Islands |url=https://doi.org/10.1007/s00024-011-0401-4 |journal=Pure and Applied Geophysics |language=en |volume=169 |issue=8 |pages=1425–1441 |doi=10.1007/s00024-011-0401-4 |bibcode=2012PApGe.169.1425E |s2cid=129573877 |issn=1420-9136}}</ref> Moreover, the GACOS project aims to refine information from HRES with [[Satellite navigation|GNSS]] zenith delays if available.<ref name=":6" />
The advantage of this method is insensitive to the presence of clouds and Global/regional/ local coverage. By using the ERA, uncertainty in ZWD in high latitudes (> 30∘), reach 1–2 cm, while uncertainty in low latitudes (< 30∘), is about 2–6 cm. Most of studies reported varying degrees of success for using this method.<ref name=":4" /> Nonetheless, ERA-interim, and ERA5 global re-analysis models are popular models provided by the European Center for Medium-Range Weather Forecasting [[European Centre for Medium-Range Weather Forecasts|(ECMWF]]) and the HRES ECMWF forecast model.<ref>{{Cite journal |last1=Eff-Darwich |first1=Antonio |last2=Pérez |first2=Juan C. |last3=Fernández |first3=José |last4=García-Lorenzo |first4=Begoña |last5=González |first5=Albano |last6=González |first6=Pablo J. |date=2012-08-01 |title=Using a Mesoscale Meteorological Model to Reduce the Effect of Tropospheric Water Vapour from DInSAR Data: A Case Study for the Island of Tenerife, Canary Islands |url=https://doi.org/10.1007/s00024-011-0401-4 |journal=Pure and Applied Geophysics |language=en |volume=169 |issue=8 |pages=1425–1441 |doi=10.1007/s00024-011-0401-4 |bibcode=2012PApGe.169.1425E |s2cid=129573877 |issn=1420-9136}}</ref> Moreover, the GACOS project aims to refine information from HRES with [[Satellite navigation|GNSS]] zenith delays if available.<ref name=":6" />
''<u>Limitation</u>'': The low spatial resolution and the original mismatch in time between the model and the SAR acquisition do not permit addressing the turbulent component that takes place at lower Spatio-temporal scales. Moreover, complex data processing can be regarded as disadvantage of this method.<ref name=":3" />
''<u>Limitation</u>'': The low spatial resolution and the original mismatch in time between the model and the SAR acquisition do not permit addressing the turbulent component that takes place at lower Spatio-temporal scales. Moreover, complex data processing can be regarded as disadvantage of this method.<ref name=":3" />
Line 182:
Line 182:
The correlation analysis between the interferometric phase and topography can recognize the amount of topography-correlated (stratified component).<ref name=":2" /> In other words, the correlation between range change in the LOS direction of InSAR and topography are related to the path length traveled by the electromagnetic wave.<ref name=":10" /> Estimation of linear and power-law relationships between the interferometric phase and the topography can be used to remove stratified parts of the troposphere. Hence, investigating the phase-elevation relationship with mathematical methods such as regression can enable us to recognize the atmospheric error in interferograms.<ref name=":5">{{Cite journal |last1=Bekaert |first1=D. P. S. |last2=Hooper |first2=A. |last3=Wright |first3=T. J. |date=February 2015 |title=A spatially variable power law tropospheric correction technique for InSAR data |url=https://onlinelibrary.wiley.com/doi/abs/10.1002/2014JB011558 |journal=Journal of Geophysical Research: Solid Earth |language=en |volume=120 |issue=2 |pages=1345–1356 |doi=10.1002/2014JB011558 |bibcode=2015JGRB..120.1345B |s2cid=53334908 |issn=2169-9313}}</ref>
The correlation analysis between the interferometric phase and topography can recognize the amount of topography-correlated (stratified component).<ref name=":2" /> In other words, the correlation between range change in the LOS direction of InSAR and topography are related to the path length traveled by the electromagnetic wave.<ref name=":10" /> Estimation of linear and power-law relationships between the interferometric phase and the topography can be used to remove stratified parts of the troposphere. Hence, investigating the phase-elevation relationship with mathematical methods such as regression can enable us to recognize the atmospheric error in interferograms.<ref name=":5">{{Cite journal |last1=Bekaert |first1=D. P. S. |last2=Hooper |first2=A. |last3=Wright |first3=T. J. |date=February 2015 |title=A spatially variable power law tropospheric correction technique for InSAR data |url=https://onlinelibrary.wiley.com/doi/abs/10.1002/2014JB011558 |journal=Journal of Geophysical Research: Solid Earth |language=en |volume=120 |issue=2 |pages=1345–1356 |doi=10.1002/2014JB011558 |bibcode=2015JGRB..120.1345B |s2cid=53334908 |issn=2169-9313}}</ref>
''<u>Limitation</u>'': The main limitation of these model is that other phase terms (e.g., turbulent atmospheric artifacts, deformation related phase, decorrelation noise) can influence the estimate of the coefficient that relates phase with elevation. This method ignores the spatial variability of tropospheric signals and can be easily biased by orbit and topographic errors.<ref name=":3" />
''<u>Limitation</u>'': The main limitation of these model is that other phase terms (e.g., turbulent atmospheric artefacts, deformation related phase, decorrelation noise) can influence the estimate of the coefficient that relates phase with elevation. This method ignores the spatial variability of tropospheric signals and can be easily biased by orbit and topographic errors.<ref name=":3" />
== Available packages for atmospheric correction ==
== Available packages for atmospheric correction ==
Atmospheric Correction for Interferometric synthetic aperture radar (InSAR) technique is a set of different methods to remove artifact displacement from an Interferogram caused by the effect of weather variables such as humidity, temperature, and pressure.[1] An interferogram is generated by processing two SAR images before and after a geophysical event like an earthquake. Interferograms are similar to a rainbow, and each colour cycle shows a specific displacement based on radar wavelength.[2]“Corrections” for atmospheric signals are an important stage of InSAR data processing in many study areas to measure surface displacment because the displacement result of InSAR can become inaccurate by 10–14 cm if relative humidity differs by 20%.[1] Overall, atmospheric correction methods can be divided into two categories: a) Using Atmospheric Phase Screen (APS) statistical properties and b) Using auxiliary (external) data such as GPS measurement, multi-spectral observations, local meteorological models, and Global Atmospheric Models.[3]
Interferometric synthetic aperture radar (InSAR) can provide accurate (millimeter-level) ground displacement fields for large areas over hundreds of kilometers.[1] This technique uses two SAR images of the same area acquired at different times to measure surface motion.[1] Nevertheless, the result of InSAR in interferogram form includes actual displacement and other effects. Hence, these other effects must be calculated and removed from interferograms to achieve an accurate result of real ground displacement.[4] Some of these errors have more influence, such as orbital errors, topographic effects and atmospheric artefacts. In addition, other noises like unwrapping and geometry errors should be considered to reach a trustable result.[1] There are several methods to remove all these noises until a reasonable and acceptable amount except atmospheric effects with high accuracy.Thus, a significant challenge for the InSAR community remains atmospheric artefacts. Sometimes especially in areas with high humidity, the effect of atmospheric noise is much larger than geophysical events and prevents surface displacement detection.[1]
On the other side, atmospheric noise might have some value for atmospheric research in Meteorology because atmospheric artefact signals are related to water vapour in the troposphere.[5] The spatial resolution of the InSAR map for C-band satellites like Sentinel-1 without multi-looking is around 20 meters. That means InSAR can measure Precipitable Water Vapor (PWV) in the atmosphere in a 20m grid over hundreds of kilometres, which is much denser than other methods such as GNSS and space-borne passive sensors.[6] However, the long revisit time of Sentinel-1 (temporal resolution, 12 days) at the moment is the main disadvantage of this technique from the meteorologists' side. Nevertheless, using the capability of InSAR to measure PWV in high spatial resolution is interesting for meteorological research.[7]
Atmospheric noise
In radar satellites, microwave signals are reflected off a persistent scatter in a target area, and their two-way travel time is measured by satellites.[1]Water vapor in the troposphere and free electrons in the ionosphere affect the propagation of microwave signals through the atmosphere because the different refractive index in these layers affects the speed of propagation.[1]
File:Satte0301.tifIonospheric and Tropospheric effects on electromagnetic wave propagation[8] This delay is caused by refraction and diffraction in different layers.
Ionosphere
Ionospheric phase noise, which occurs more apparent with larger wavelengths, such as P or L-band radars, is a consequence of variations in free electron density in the 100–1000 km altitude along the travel path.[9] Radar satellites with large wavelengths like ALOS Polsar1/2 (L-band, λ=~24 cm) and NISAR are more vulnerable to ionospheric delays. However, this noise is roughly less effective in C-Band (λ=5.6 cm) and X-band (λ=3.1 cm) SAR satellites such as Envisat, RADARSAT, ERS1/2 and TerraSAR-X and usually negligible.[9] There are a couple of methods to remove the Ionospheric noise artifact:
File:Tenn09.tifAn example of atmospheric artifacts in the interferogram of Tenerife, Spain.[14] This image shows the correlation between height and displacement clearly.
Troposphere
The troposphere is the lower layer of the atmosphere, with up to 90% of the atmosphere's water vapor. A height of 1.4 km from the ground surface contains 50% of the water vapor mass, and both on the global and local scales, most of the world's weather takes place in this layer.[6] Therefore, the tropospheric path delays are caused by differences in temperature, atmosphericpressure, and the water vapor in the lower part of the atmosphere. Microwave atmospheric delay is a sum of the turbulent and vertical stratified components. The turbulent component can affect flat and mountainous terrains and is usually highly correlated in space and uncorrelated in time. On the other hand, the vertical stratified component with a high correlation to topography only influences areas with topography variations, such as hills or mountains.[1][15]
Although the tropospheric effect is regarded as noise in the InSAR community, it has great advantages in meteorology and enables scientists to predict and model water vapor in the troposphere.[5]
The propagationtion delay of electromagnetic waves of radar satellites in the troposphere will be explained in the following in terms of meteorology and the InSAR side.[1]
Layers of troposphere and amount of water vapor in each altitudeAmount of water vapor in each altitude of the troposphere. Around 90% of water vapor is contained in the troposphere and 50% of which is between the earth's surface up to about a height of 1.4 km
From the InSAR side, signal delay caused by variations of tropospheric properties in space and time is the source of major challenges for InSAR technique. In other words, tropospheric perturbation, caused by the differences in relative humidity, temperature, and pressure in the lower part of the troposphere between two acquisitions, can lead to additional noise in form of fringes up to 15–20 cm on interferograms.[16] The atmospheric noise on InSAR results can include a wide range of wavelengths (short to long). Long wavelength errors, usually seen as a ramp (like a trend) in interferograms, are caused by changing the weather system in the study area very smoothly between two SAR images. Since this noise is similar to orbital ramp error and solid earth tides, detecting that in an interferogram is complicated.[17] On the other hand, rapidly changing weather in a small area can cause artefact signals that correlate with topography because water vapour variation in surface and altitudes is different. Moreover, a rain cloud in a small region can generate a turbulence error which would be visible like uplift or subsidence on interferogram.[18] Overall, the tropospheric error on interferogram can be classified into space and time[18]:
Tropospheric artifact
Space
Systematic
Stratified
Stochastic
Turbulent
Time
Systematic
Seasonal
Stochastic
Nonseasonal
File:Reflection0010013s.tifThese figures show the correlation between the signal delay and altitude in two points( p and q ) with different hight at t1(date of image one) and t2(date of image two). The signal reflected from q would have more delay because the amount of water vapor near the surface is very high and subject to more variation in t2.[19]
Vertical stratification: Vertical stratification is correlated with topography and can be seen wherever there are hills or mountains because the amount of water vapour is not constant near the surface and changes significantly with altitude. Hence, changing weather (water vapour) between two SAR images can cause delay and, thus, vertical stratification noise in the interferogram.[19]
Turbulence: Turbulent mixing consists of a short-scale (few km) components and is not directly correlated with topography and affects both flat terrains and higher altitude terrains.[19]
File:Teneaw.jpgPopocatepetel, Mexico, 2014-2017 summit displacement and seasonal effect during 2016 (Adopted from Falk Amelung presentation, InSAR Meteorology Miami 2018[5]).Seasonal (systematic) effect : Variation of moisture content of the atmosphere during different seasons may impose an artifact and bias the InSAR displacement time series. The large number of SAR acquisitions will not reduce the velocity bias necessarily.[18] Nonetheless, the stratified tropospheric delay correction can mitigate the effects of the seasonal delay.[18][20]
In terms of Meteorology, tropospheric delay interestingly can be regarded as a useful tool for Meteorology purposes.[5] Traditional methods for measuring water vapor include using a) Meteorological stations, b) Radiosonde, c) Spectrometer, and d) GPS.In recent years, InSAR has enabled meteorological scientists to measure precipitable water vapour (PWV) in the atmosphere with high spatial resolution.[6] Although InSAR temporal sampling (12 days for sentinel1) of atmosphere properties is coarser than the GPS (15 min), the main advantage of InSAR PWV maps is its high spatial resolution of up to a few meters (20 m).[6] Therefore, meteorological scientists combine traditional GPS tomography of atmosphere with InSAR data (atmospheric part) to increase their models' temporal and spatial resolution.[21]
File:Ppt file Seminar 28.jpgSchematic resolution comparison of InSAR and GPS for a large area. Red points in InSAR and GPS show pixel resolution, and in GPS stations, respectively.
The atmospheric parameters can be used to calculate N(z) as the atmosphere's refractivity.[22] This delay can be divided into two components: Dry delay and Wet delay. The dry delay, which is determined by the temperature and pressure of dry air, and the wet delay, which is determined mainly by the amount of water vapor in the troposphere.[6]
where P indicates total atmospheric pressure, T is the temperature and e is the partial pressure of water vapor. The coefficients k1 =0.776 K/Pa, k2 = 0.716 K/Pa, and k3 = 3.75 × 103 K2/Pa are called refractivity constants.[20][23]
Since pressure varies with height, the dry part is correlated with the topography and can reach a 2.3–2.4 m delay in the zenith direction (zenith hydrostatic delay (ZHD)). Although the dry part significantly affects delay, this component is not a big challenge because the amount of that is relatively stable in terms of temporal and spatial in the atmosphere. Therefore, the differential of this component between two SAR images is almost zero. On the contrary, although the wet component (zenith wet delay (ZWD)) includes just 10% of the total delay (max =~30 cm ), this part is the major source of tropospheric noise in InSAR because water content moves in the troposphere and cause stratified and turbulent effect in the interferogram. [6]
Therefore, the total delay is calculated by the integral of the wet and dry components in the line of sight (LOS) direction between the surface elevation and satellite:[22]
Weighted average of the gravity acceleration between z and zre
Rd
dry air specific gas constants = 287.05 J.kg−1.K−1
e
Water vapor partial pressure in Pa
Rv
Water vapor specific gas constants = 461.495 J.kg−1.K−1
p
Dry air partial pressure in Pa
T
Temperature in K
Since there is no accurate technique that could determine the total refractivity on the same spatial scale and temporal sampling as the interferogram itself, various methods have been developed to measure for the atmospheric contribution within interferograms. In the following a couple of methods are introduced to mitigate tropospheric artifacts from interferograms:
Time series methods such as Persistent Scatter (PS) and Small Baseline Subset (SBAS)
Stacking has been one of the helpful strategies for reducing the tropospheric artefacts from interferograms.[18] The stacking method processes several to hundreds of filtered and unwrapped interferograms together using different algorithms like Small Baseline Subset (SBAS) and persistent scatterer (PS) to measure surface displacement during a long period.[24][25] These methods increase the signal-to-noise ratio of signals without any external information and are very useful for mitigating turbulence errors.[18][26][27] The other advantage of this method is independent on external data and straightforward to implement.
Limitation: For time series analysis approaches such as PS and SBAS, a large dataset has to be processed and this method mostly can reduce turbulent part of troposphere noise and signals of interest may be incorrectly removed. Thus, this method cannot be implemented only for one interferogram.[18]
Estimated tropospheric delay from GNSS
GNSS uses microwave portion of the electromagnetic spectrum similar to SAR technique and both measurements are affected by the atmospheric delay and both provide geodetic measurements with comparable accuracy. Therefore, using interpolating GNSS observation (ِdense GNSS network) to estimate tropospheric delay can be a more accurate strategy to correct InSAR observations.[28]
Limitation: Interpolation of zenith delay measurements requires dense GNSS network in the study area but GNSS stations are still sparsely distributed or even absent in many regions as well as this method only samples troposphere in the vicinity of individual GNSS sites. Moreover, GNSS datasets are still not freely available to the public in the many areas in the world.[29]
Spectrometer measurements are Satellite-based observations from space allow estimating the atmospheric water vapor by band ratios in the near-infrared spectrum. This method has a relatively high spatial resolution to calculate the turbulent component of the atmospheric disturbance.[30]
Limitation: Requires collocated sensors and cloud-free conditions and only available in daytime. Time differences between radar and Precipitable Water Vapor (PWV) data can be regarded as limitation.[16] Moreover, the Spectrometer cannot estimate the atmosphere's dry component and calculate the wet part of the delay.[16]
Numerical weather models (ERA5)
The advantage of this method is insensitive to the presence of clouds and Global/regional/ local coverage. By using the ERA, uncertainty in ZWD in high latitudes (> 30∘), reach 1–2 cm, while uncertainty in low latitudes (< 30∘), is about 2–6 cm. Most of studies reported varying degrees of success for using this method.[20] Nonetheless, ERA-interim, and ERA5 global re-analysis models are popular models provided by the European Center for Medium-Range Weather Forecasting (ECMWF) and the HRES ECMWF forecast model.[31] Moreover, the GACOS project aims to refine information from HRES with GNSS zenith delays if available.[32]
Limitation: The low spatial resolution and the original mismatch in time between the model and the SAR acquisition do not permit addressing the turbulent component that takes place at lower Spatio-temporal scales. Moreover, complex data processing can be regarded as disadvantage of this method.[18]
Freely available Global Atmospheric Models such as ERA5 data are generated using Copernicus Climate Change Service Information developed for numerical weather prediction. These models can provide accurate and high-resolution (still is coarse for InSAR studies) parameters for characterizing the atmosphere state. Therefore, these methods are promising and practical techniques for atmospheric noise mitigation in InSAR technique. Based on the combination of different input datasets, the ERA5 is a global atmospheric model calculated by the European Center for Medium-Range Weather Forecasting (ECMWF). Several meteorological parameters are provided, such as pressure, temperature, and relative humidity, at hourly intervals with a horizontal resolution of 0.25 degrees and a vertical resolution of 37 intervals from sea level to 50 km.[20]
This method estimates the atmospheric delay along the zenith path (vertical), and then, by using the incidence angle for each pixel, the zenith delay is converted to LOS direction.[20] In order to obtain the phase delays for the entire SAR scene from the sparse grid points, two interpolations are implemented in horizontal and vertical directions. ERA5 data provides all the weather variables at 37 pressure levels (geopotential). First, the amount of delay is calculated in the signal path using cubic spline interpolation (vertical direction) and bilinear interpolation (horizontal direction). Then the total delay is projected to the LOS direction by calculating the cosine of the incidence angle for each pixel.[20] Thus, after calculation in first image t1 and second image t2, and subtraction of the original interferogram (), the effect of atmospheric noise can be removed.[20]
Atmospheric correction across the Tenerife. The original interferogram contains atmospheric noise; by using ERA5 data can remove the atmospheric artefact. The residual shows the interferogram without atmospheric noise.[3]
It is worth mentioning that, the equation (see previous section) is an integration of zenithal path on the grid points of Pressure, Temperature and relative humidity to measure atmospheric phase delays. These paramaters are available in the ERA5 data:
Phase based empirical: (i) Linear correction (ii) Power-law correction
The correlation analysis between the interferometric phase and topography can recognize the amount of topography-correlated (stratified component).[16] In other words, the correlation between range change in the LOS direction of InSAR and topography are related to the path length traveled by the electromagnetic wave.[19] Estimation of linear and power-law relationships between the interferometric phase and the topography can be used to remove stratified parts of the troposphere. Hence, investigating the phase-elevation relationship with mathematical methods such as regression can enable us to recognize the atmospheric error in interferograms.[33]
Limitation: The main limitation of these model is that other phase terms (e.g., turbulent atmospheric artefacts, deformation related phase, decorrelation noise) can influence the estimate of the coefficient that relates phase with elevation. This method ignores the spatial variability of tropospheric signals and can be easily biased by orbit and topographic errors.[18]
Available packages for atmospheric correction
Terrain: TRAIN consists of MATLAB and shell scripts and can be used for the output of most InSAR software. [33]. (click on the link )
PyAPS: This module has been written based on python 3 and can measure the stratified atmosphere noise for interferograms. This module uses ERA5 data for correction.[20] (click on the link )
RAiDER: This python package implements the raytracing method to measure and reduce tropospheric noise. .[34] (click on the link )
GACOS: This package in MATLAB generates high-resolution atmospheric data and then separates stratified and turbulent signals to remove tropospheric noise. This method uses the Iterative Tropospheric Decomposition (ITD) model.[35][36][32] (click on the link )
ICAMS: This python module uses ERA5 data and considers stochastic spatial properties of the troposphere (ICAMS) to remove tropospheric noise.This package calculates the delays along the LOS direction.[37] (click on the link )
Summary
Characterizing atmospheric noise remains a challenge in the InSAR community, and addressing it helps researchers to take full advantage of the InSAR technique.[37]
All methods to mitigate this noise have limitations; sometimes, combining techniques gives a better result, and there is no best exclusive method for reducing tropospheric delays at the moment.[29]
Global measurements of tectonic/volcanic deformation commonly benfit from global atmospheric corrections. Although ECMWF data, like ERA5, provides global data, the low spatial resolution of it regarded as a drawback and can cause uncertainty.[29]
This article has not been added to any content categories. Please help out by adding categories to it so that it can be listed with similar articles. (November 2022)