Satellite 101, part 1 - Transcript

Satellite 101 Part 1

 

Hello! Welcome to Part 1 of Satellite 101, part of an online course on oceanographic  satellite data products, which has been produced by NOAA's CoastWatch Program. My name is Cara Wilson, I'm the node manager of  the West Coast Node of NOAA's CoastWatch Program. The materials I will present in this video  were produced from a collaboration from many members of NOAA's CoastWatch Program  including Dale Robinson, Melanie Abecassis, Ron Vogel, Shelly Tomlinson,  myself, and the late Dave Foley.

 

In part 1 of the Satellite 101 presentation, I will introduce some of the basic terminology used in the field of remote sensing. I will give  an overview of available oceanographic satellite products and present the different types of  orbits, resolutions and spatial coverages. I will also briefly go over the  different levels of data processing. In part 2 of the Satellite 101 presentation, I will introduce the different types of satellite 

sensors and how measurements are made. There  are also additional presentations as part of the CoastWatch course that go into more detail about  how certain measurements and products are made.

 

I like to use a food analogy to explain the approach of this course. I'm not a vegetarian and I like sausages, but I'd prefer to not know what  goes into them, it's often a lot of nasty body parts! And I feel the same way about satellite  data. I just want to use the data without having  to learn about all the processes that go into  creating these products, information about  electromagnetic radiation, wavelength bands, IOPs and AOPs, atmospheric correction, lunar calibration etc. etc. But just as one has to make  some decisions when buying sausage at the store; chicken or pork, or tofu for my vegetarian  friends, nitrate-free or not, one has to know a little bit about satellite products when  using them. In this course we aim to give you the basic information you need to make decisions  about choosing a satellite product to use. This course does not intend to teach the  sausage making components of satellite data. That amount of detail needs a much  longer and comprehensive course.

 

So what measurements of the ocean can we make from satellites? We can measure the amount and type of sea  ice, we can measure sea-surface temperature, or SST for short. We can measure sea-surface  height of the ocean, or SSH for short. By measuring ocean color we can derive the  chlorophyll concentration of the surface ocean. We can also measure the speed and direction  of ocean currents and winds over the ocean, the amount of rainfall over the ocean,  as well as the sea surface salinity. For how long have we been making these measurements?

 

The answer depends on the measurement. On this slide the different measurements are arranged in order of the  length of the timeseries. Measurements of sea ice have been collected  the longest, with the continuous record going back to 1978. The salinity is a relatively new  measurement, which started in 2011. Note that the years shown here indicate the start of the  continuous record of each of these parameters, but in some cases the first measurements were made earlier. For example the first ocean color satellite was the CZCS, the Coastal  Zone Color Scanner, which flew from 1979 to 1986, but there was a ten year gap between that  data and the launch of SeaWiFS in 1997, which marks the beginning of the continuous  record of ocean color measurements.

 

This slide shows a little bit  more graphically the progression of different oceanographic satellite  missions over the last few decades. Not only has the number of available products  been growing, but also the number of satellites, and now there can be multiple  satellites measuring the same parameter. Many countries have satellite programs, and there  is international cooperation and data sharing.

 

So why do we need satellite data? The  oceans are huge, covering 71% of our planet, and it is impossible to adequately  measure them with in-situ measurements. Satellite data is able to provide us measurements of the ocean at temporal and spatial scales that we could never cover with ships or buoys, alone. The image on the left shows a compilation of all the in-situ sea-surface temperature data  collected in a month, and the image on the right shows a compilation of all the satellite  sea-surface temperature data collected in one day! Having continuous satellite measurements makes it  possible to detect anomalous events and also to go back and observe past events using the archived  record of measurements. An important caveat though is that the satellite measurements are just of the  surface of the ocean, so for information on the subsurface structure of the ocean we need to rely  on a combination of in-situ data and ocean models.

 

There are two main types of satellite orbits: polar-orbiting and geostationary. The animation on the left depicts a polar-orbiting  satellite. As the name implies the satellite orbits around the poles of the earth, as the earth  spins, the satellite measures a swath below it. In the course of one to several days the  satellite will have passed over the entire globe. Geostationary satellites have orbits  that are in sync with the Earth's orbit, so they are always looking at the same area of  the Earth, allowing frequent measurements over the region instead of only one measurement per  day. Geostationary satellites sit a lot farther away from the earth than polar-orbiting satellites  and therefore can observe much more surface area. Most of our environmental data comes from  polar-orbiting satellites, and most of our weather data comes from geostationary  satellites. Some SST measurements are made from geostationary orbit, and South Korea has an  ocean color sensor on a geostationary satellite.

 

This image shows all the data collected in one day  by the polar-orbiting VIIRS ocean color sensor on the NOAA-20 satellite. The black areas are regions  where no measurements were made. The streaks running north to south through the image are areas  where the satellite did not pass over. The other black regions are regions where measurements  were not made because of cloud coverage.

 

While there are some areas on the globe that will  be missed each day by a polar-orbiting satellite, because each pass goes over the poles, they  collect a lot of data at high latitudes. This image shows the 14 passes made in one  day of the SNPP satellite over the north pole.

 

Here we can see the global coverage of  geostationary satellites. Because each satellite can only observe part of the planet, multiple  satellites are needed to provide global coverage. We need two satellites to adequately cover the  United States, we have GOES-West and GOES-East. Those satellites, combined with  the two European satellites and the Japanese geostationary satellite,  can provide nearly global coverage. However even the two US geostationary  satellites do not cover all of the United States. Their coverage does not extend  into high latitudes and unfortunately we can’t currently make geostationary  measurements over much of Alaska.

 

This slide summarizes the difference between  polar-orbiting satellites and geostationary satellites. Polar-orbiting satellites provide  global coverage, at relatively high spatial resolution, on the order of 1 km resolution, but  a low temporal resolution, on the order of 1 or 2  days. In contrast, geostationary satellites provide  regional coverage, at lower spatial resolution,  on the order of 2-4 km, but high temporal  resolution, with approximately hourly resolution. With satellite data there is always a  trade-off when it comes to resolution: higher temporal resolution means lower spatial  resolution, and vice-versa. You can't have both!

 

I've mentioned the different  types of resolution of satellites,  let's talk about that a little more. Resolution  of a satellite can refer to a number of things. The spatial resolution is the smallest geographic  area that a satellite instrument can resolve. The spatial resolution of oceanographic  satellite products ranges from 250 m to 25 km. The higher the spatial resolution,  the more details that can be observed. The Temporal resolution is how frequently a  satellite observes the same area on Earth. The Spectral resolution refers to how many bands the  sensor has. The higher the spectral resolution, the more specific types of observations  are possible with the same instrument.  Another relevant specification is the Swath  width, which refers to the width path observed by a polar-orbiting satellite as it orbits  that Earth. Satellites with larger swath widths typically have a greater temporal  resolution and a lower spatial resolution.

 

There are other satellites that  have higher spatial resolutions, on the order of 1 to 30 meters, compared to the 1-25 km we deal with in ocean remote sensing. I previously mentioned the trade-offs inherent  with different resolutions. These very high spatial resolution sensors can have very long  repeat times, or might not have any repeat times at all, but rather work on on-demand acquisition.  These data can be challenging to obtain and cumbersome to process, and often better suited  for land applications than for ocean applications.

 

Although these high resolution data are not  generally offered as part of this course, I will show you a cool example of a study  that used one high resolution dataset for an oceanographic application. This study used  data from the WorldView sensor, which has a 50 cm resolution, to try to identify whales off of Peninsula Valdes in Argentina. The white objects in the images are probable  whales that were identified by automated analysis.

 

A lot of processing goes into satellite data  before reaching you, the sausage making that I  referred to earlier. There are terms to refer  to how processed the data are. For example Level 0 data is the raw data received from  the satellite, level 1 data has been put into the satellite s geographical coordinates  (i.e along-track ) , level 2 data has been processed into a derived geophysical variable,  like sea surface temperature, but is still in swath coordinates. Level 3 has been mapped onto  a standard map grid with latitude and longitude. Level 4 data has had an additional degree of  analysis performed, perhaps deriving another product, performing interpolation to remove  gaps, temporal compositing etc.This course focuses on accessing and using level 3 and  level 4 data, i.e. off the shelf data products.

 

Let's talk briefly about temporal composites. Many  satellite products contain gaps where there are no measurements due to cloud coverage. These gaps  can be at least partially filled in by making   composites, or averages of different satellite  images. There are two different ways this can be done. Temporal composites of a sensor can be  made over longer time periods. Most satellite products are composited into daily, weekly and  monthly products. For measurements that are made by multiple satellites, blended products can be  made that merge data from the different sensors.

 

This slide shows an example of 4 different  temporal composites of geostationary sea-surface temperature data in the Pacific Ocean. The  image on the top left displays one hour of data from Sept 15, 2018. There are more areas  with missing data than there are areas with data. The image on the bottom left shows a daily  composite of data from this same sensor, and most of the missing data from  the image above it are now filled in. The image on the top right shows a weekly composite of the same area. Now there is only a small region with no data,  around a longitude of 230 , where it has been cloudy all week. The monthly composite in the  bottom right is gap-free, except for the region in the northwest part of the map, which  is outside of the disk view of the satellite.

 

The composites we've just seen are necessary to  make because of data gaps due to cloud coverage. The cloud issue is a huge problem for ocean remote  sensing. But of course the satellite doesn't stop making measurements when there are clouds, but  rather it makes measurements of the cloud, not the ocean below it. So in order to remove this cloud  data we need cloud masks that determine what data is from a cloud and what is from the ocean. Cloud  masks are usually made from visible imagery, so nighttime retrievals of SST can be less accurate,  since the cloud masks might not be as reliable. Different agencies and different satellite  products can use different masks.

 

For many applications it is more  useful to view how a parameter  differs from a typical or mean value. This is  called an anomaly product. For example, to see the extent of the recent marine heat wave  in the North Pacific, one needs to look at the Sea Surface Temperature anomaly, rather  than Sea Surface Temperature. To generate an anomaly one needs a climatology mean of the  parameter to subtract from the timeseries. Here I have listed some of the  anomaly products served by CoastWatch.

 

It's important to realize the distinction  between satellites and sensors. Sensors are the instruments making the measurements. Satellites are the platforms carrying the sensors. Currently most satellites have many different  sensors on them. This slide shows all of the sensors on the JPSS-1 (NOAA-20) satellite.  One of these sensors is VIIRS, a sensor  which is also on the SNPP satellite which  was launched a number of years before JPSS-1.

 

In closing I'll mention the two major  satellite agencies in the US, NASA and NOAA. Both agencies have satellite missions but they  have different responsibilities and objectives. NASA is responsible for research and development  of satellite instruments and products, whereas NOAA's role is to provide operational  satellite data for routine use. Those are two very  different objectives. NOAA needs to measure and  expand how its satellite products are being used operationally, in order to ensure a continuous and  reliable suite of environmental measurements. And that need is indeed what led to the development  of the NOAA CoastWatch satellite course many years ago. Oceanographic satellite data was being  underutilized within the wet part of NOAA, i.e. NMFS and NOS, and this course was created to  increase the utilization of satellite data within  NOAA, and in other operational applications. More details on how satellite measurements are made are presented in Part 2  of the Satellite 101 presentations