Surface Water and Remote Sensing Notes

Introduction

  • Introduction of Participants

    • Yunsang and Sarah
    • Adriana (course GE)
    • The class

  • Weather Discussion

    • Palo Alto: 68 and sunny, 75 and sunny tomorrow.
    • Eugene: Rainy stretch, but nice snow in the mountains.

  • Guest Speaker Introduction

    • Sarah is the guest speaker.
    • The class has reviewed her article and research website.

  • Topic Introduction

    • Thematic application area: Water and ice (cryosphere and hydrosphere).
    • Sarah will focus more on the hydrology side.

  • Sarah's Background

    • Currently a postdoc at Stanford.
    • Starting as a professor at UO in the geography department in September.
    • Topic: Observing surface water from space.

Understanding Global Hydrology

  • Global Hydrology

    • Understanding the water cycle at a global scale.
    • Includes precipitation, groundwater, ice, snow, and glaciers.
    • Focus on surface water: rivers and freshwater lakes.

  • Importance of Understanding Global Surface Water

    • Better understanding global water resources.
      • Knowing where water is and how much is available. (e.g., Hetch Hetchy reservoir in Yosemite).
    • Understanding flood events.
      • Mapping flood events helps predict future occurrences.
    • Understanding human management.
      • Humans exert a strong role in regulating surface water via dams.
    • Studying climate change.
      • Remote sensing helps monitor the effects of climate change on the water cycle.

  • Ways to Study Global Hydrology

    • Stream and lake gauges.
      • USGS river gauging station in Alaska.
      • Simple gauges (stick in a river or lake).
    • Fieldwork.
      • Measuring discharge of a small stream on the Greenland ice sheet.
      • Surveying a small lake in Northern Alaska using GPS.
    • Limitations of in situ approaches:
      • Expensive and hard to travel to many places.
      • Gauges are limited; data is often unavailable, particularly in Asia and Africa.
      • Even fewer lakes are gauged.

  • Remote Sensing

    • Becoming the dominant tool for understanding global hydrology.
    • 40-year records of satellite imagery available.
    • Example: Landsat image of Lake Powell.

Case Studies

  • How has surface water area changed over the past forty years?
  • How does the amount of water stored in lakes, ponds, and reservoirs vary seasonally?

How Has Surface Water Area Changed Over the Past Forty Years

  • Water Classification in Satellite Imagery

    • Water is one of the easiest things to classify because of its distinctive spectral response curve.
    • Water reflects light in the visible spectrum but absorbs light in the near-infrared spectrum.
    • Water usually appears dark in satellite imagery, whether visible or false color composites.
    • Exceptions: water may appear lighter due to algal activity or sediment.

  • Spectral Indices

    • Normalized Difference Water Index (NDWI) is commonly used.
    • NDWI=Green−NIRGreen+NIRNDWI = \frac{Green - NIR}{Green + NIR}
    • NDWI uses the largest differences between green and near-infrared reflectance.
    • Example: Landsat composite of the Eugene area and corresponding NDWI image.
      • The water appears very bright in the NDWI image.

  • Surface Water Mapping

    • Many studies have used this approach to classify surface water over small areas.
    • Example: Studying how lakes are changing using high-resolution cubesat imagery.

  • Key Advances for Global Mapping

    • USGS made the Landsat record free in 2008.
      • Before, a Landsat image could cost $600-$4,000.
    • Advances in computing and cloud technology.
      • Google Earth Engine allows analyzing large volumes of imagery.

Global High Resolution Map

  • Pekel et al. 2016 Paper

    • Classified water over the entire archive of Landsat (Landsat 4, Landsat 7, and Landsat 8).
    • 1,800 terabytes of Landsat data spanning 1984 to 2015.

  • Number of Images

    • Highlights that prior to, images over America were more abundant, because of testing happening there.
    • Highlights the effect of cloud cover, less imagery is available along the equator compared to surrounding regions as the ITCZ zone is much cloundier than surrounding regions.

  • Data Availability Over Time

    • Shows the launch of Landsat 7 and Landsat 8 increased the number of images available over time.
    • Landsat 6 failed to properly reach orbit.

  • Expert Classifier

    • Classified images into three core categories: water, land, or nonvalid.
    • Nonvalid includes cloudy, snowy, and other things.
    • Expert classifier uses a series of if-then statements.
    • Challenges: How to remove other things that are not water but also have spectral signatures that look like water.
      • Terrain shadows
      • Glaciers
      • Buildings
      • Lava
      • Cloud shadows

  • Global High Resolution Map of Surface Water

    • For every pixel, calculated the percent of time that pixel is water and use it to assign whether it's permanent water or seasonal water.
      • A lot of water in the northern regions and far less water south of the Equator.
    • It's the highest resolution (30 meter resolution) map of where water occurs globally.
    • Created over a long record (forty years).

  • Global High Resolution Map of Trends in Surface Water

    • Identifies areas where some of the most dramatic changes in surface water are occurring.
    • Green areas are where there is more water, and red/blue areas are where there is less water.
    • Increases in water are seen in Canada and the Eastern US.
    • Losses of water occur in parts of the Western US.

  • Aral Sea

    • Endorheic basin in Central Asia.
    • Shrunk dramatically because Soviets built dams and rerouted rivers for irrigation.
    • Total ecosystem collapse; environmental disaster.

  • Tibetan Plateau

    • Increases in surface water because of glacial melt and changes in snowmelt.

  • Southeast Asia

    • Extensive dam construction over the past forty years, many by Chinese government.
    • Myanmar has a lot of new dams and lakes that were created over this period.

  • Conclusion

    • The dataset is fully accessible online.
    • Enabled uniquely by Google Earth Engine and Landsat technology.

A Different Kind

  • Area vs Volume

    • Mapping water area is not the same as mapping water volume

  • Satellite Altimetry

    • Needed to measure water height, unless you have a gauge.
    • Traditionally has worked through satellite radar altimeters.
      • Mostly used to map ocean height.
    • TOPEX Poseidon, Jason satellite series used to map the ocean.

  • Monitoring Lakes

    • USDA GRALM Dataset is most accurate in monitoring lake height, using satellite technology.

  • ICESat-2

    • Launched in October 2018, intended for studying primarily ice sheets.
    • Carries a laser altimeter (photon pulse).
    • Has six different beams that point photons at the Earth.
    • Repeat cycle of ninety days.
    • Observations are accurate to less than five centimeters

  • Example

    • Observations directly intersect the lake.
    • Details are useful in assessing the water levels.
    • Great agreement is found between ICESat-2 and the ground gauges.

Pretty High Variability

  • Intersecting ICESat-2

    • Intersecting ICESat Two with dataset, produce datasets for lake and pond water levels.
    • Can look at areas smaller than before because dataset is so high resolution. (point zero one square kilometers)

  • Patterns of Observation

    • Waters observe in the tropics have high variability in dry season.
    • Seasonal is the highest vs the lowest water level that can be viewed through observation of location.

  • Human Management

    • Humans heavily changing the globe water cycle.
    • 57% of variability of water can be controlled by humans.

  • Reservoirs

    • Allow irrigation.
    • Produce hydropower.
    • Threat: Salmon in the pacific northwest and downstream ecosystem.

  • South Africa

    • Ran out of water a few years ago.
    • Fine like in water measurement

The Greenland Water

  • Rating Curves

    • By observations, they can produce storage time series.
    • Combine observations of area from ICESat 2 with observations of level from Sentinel 2 to produce a time series of reservoir storage.

  • The Future of Water Observation

    • Surface Ocean topography mission, scheduled to launch in 2022.
    • This new satellite technology is at a higher resolution to better observe the lakes and rivers in the water system of the globe.

Conclusion

  • Takeaways

    • Major use of remote sensing, clouds allowed mapping.

  • New Advents

    • New technology offers multiple new ways and opportunities in mapping surface water. (ICESAT. NISAR, Sentinel, Cub Sats etc.)