Remote Sensing of Lake Dynamics in Remote Sensing of Lake Dynamics - - PowerPoint PPT Presentation

▶

Nov 21, 2023 9 likes •423 views

University of Texas at San Antonio October 19, 2007 Remote Sensing of Lake Dynamics in Remote Sensing of Lake Dynamics in the Context of Global Change: A the Context of Global Change: A Global Perspective Global Perspective Yongwei Sheng

SLIDE 1

October 19, 2007 University of Texas at San Antonio

Remote Sensing of Lake Dynamics in Remote Sensing of Lake Dynamics in the Context of Global Change: A the Context of Global Change: A Global Perspective Global Perspective Yongwei Sheng

UCLA Department of Geography

SLIDE 2

October 19, 2007 University of Texas at San Antonio

Global Lake Distribution from GLWD Global Lake Distribution from GLWD

~250,000 lakes (>0.1 km2) Largest group of lakes:

high-latitudes (> 45oN);

Second largest:

27 -- 42oN;

… …

Source: GLWD (Lehner and Doll, 2004)

Complied from various sources (1 : 1 to 3M):

DCW (1970s to 1990s); Arc World (1992); WCMC Wetlands map -- World Conservation Monitoring Center (1993)

Currently best available data sets.

SLIDE 3

October 19, 2007 University of Texas at San Antonio

Problems of GLWD Problems of GLWD

(Walter et al., 2007)

A good reference:

250k lakes; 2.4 million km2; 1.8% density.

Miss a lot of small lakes: Not a systematic inventory; Not addressing lake dynamics. Another global lake estimate (Downing et al, 2006):

Lake abundance: >300 million lakes; Total lake area: 4.6 million km2; Lake area density: >3%.

SLIDE 4

October 19, 2007 University of Texas at San Antonio

Lake Dynamics Lake Dynamics

Water & energy cycling; “measure, monitor, and forecast the US and global supplies of fresh water.” (OSTP, 2004) Global warming:

How much have lakes changed? What are the mechanisms? What are the possible consequences?

But, How?

Remote Sensing!

SLIDE 5

October 19, 2007 University of Texas at San Antonio

Our Current Critical Regions for Our Current Critical Regions for Lake Dynamics Remote Sensing Lake Dynamics Remote Sensing

West Siberia (~ 0.5 M km2); Pan-Arctic (~ 40 M km2); Tibetan Plateau (~ 1.5 M km2); All remote, under-populated, climate-sensitive. .

SLIDE 6

October 19, 2007 University of Texas at San Antonio

Local-Scale Arctic Lake Dynamics Local-Scale Arctic Lake Dynamics

Studies have recently used remote sensing, field work, and historical records to examine Arctic/sub-Arctic lakes changes during recent decades:

Osterkamp et al., 2000 Jorgenson et al., 2001 Yoshikawa and Hinzman, 2003 Christensen et al., 2004 Payette et al., 2004 Stow et al., 2004 Marsh et al 2005.

Most of them are done at local scale. Does lake dynamics exhibit a pattern?

SLIDE 7

October 19, 2007 University of Texas at San Antonio

Regional-scale Regional-scale Lake Dynamics Lake Dynamics in West Siberia in West Siberia

Satellite-based inventory of an area > 0.5 million km2. 1973 MSS imagery

vs. 1997/98 RESURS

imagery.

SLIDE 8

October 19, 2007 University of Texas at San Antonio

Inventory of ~10,000 large Siberian lakes (1973-1998) reveals lake growth in continuous permafrost but disappearance in discontinuous, isolated and sporadic permafrost

(“Disappearing Arctic Lakes,” Smith et al., Science, 2005)

SLIDE 9

October 19, 2007 University of Texas at San Antonio

lake shrinking further south lake expansion (northern, continuous permafrost)

Ground Ground Confirmation Confirmation

SLIDE 10

October 19, 2007 University of Texas at San Antonio

http://earthobservatory.nasa.gov/Newsroom

125 disappeared lakes (> 0.4 km2) were detected! No new lakes.

SLIDE 11

October 19, 2007 University of Texas at San Antonio

Mechanism for Arctic Mechanism for Arctic Disappearing Lakes Disappearing Lakes

SLIDE 12

October 19, 2007 University of Texas at San Antonio

Remote Sensing of Pan-Arctic Lakes Remote Sensing of Pan-Arctic Lakes

~200,000 lakes (sized 0.1 – 50 km2, GLWD) northwards

f 45oN

Regional scale studies:

5,400 km2 lake change detection in western Arctic coast of Canada (Marsh et al, 2005); 34,570 km2 lake mapping in North Slope of Alaska (Frohn, Hinkel et al, 2005; Hinkel et al, 2007); 515,000 km2 lake change detection in West Siberia (Smith, Sheng et al., 2005).

SLIDE 13

October 19, 2007 University of Texas at San Antonio

Lake Changes at Pan-Arctic Scale Lake Changes at Pan-Arctic Scale

45oN and north:

Peak in the global lake distribution; 45oN: about the southern limit of permafrost zones; Coverage:

73 million km2, 1/7 of the Earth’s surface; 41 million km2 of land, ~1/4 of the Earth’s land surface.

So far only <2.5% of the high-latitude land surface has been studied for lake change-detection. “How have northern lakes responded to rising Arctic temperatures?” Arctic lake changes would have significant ramifications for hydrology, ecology, carbon cycle, and so on.

SLIDE 14

October 19, 2007 University of Texas at San Antonio

Requirements to Arctic Lake Requirements to Arctic Lake Remote Sensing Remote Sensing

Characteristics of Arctic lakes:

Abundant in number; Small in size; Shallow in depth; Frozen most of the time; Low-relief landscapes.

Requirements to remote sensing:

Many, high-resolution, summer images! Pan-Arctic lake mapping requires ~1,800 scenes of cloud- free Landsat images acquired in summer season for each change detection episode. Automation!

SLIDE 15

October 19, 2007 University of Texas at San Antonio

Critical Technologies and Critical Technologies and Automation Automation

Precise image co-registration; Accurate lake mapping; Detailed change detection.

SLIDE 16

October 19, 2007 University of Texas at San Antonio

Hierarchical Lake Mapping Hierarchical Lake Mapping

Global segmentation and local adjustment

SLIDE 17

October 19, 2007 University of Texas at San Antonio

Automated PIF-based Image Co-registration Automated PIF-based Image Co-registration

PIF: pseudo invariant features MSS: June vs. August, 1973

C1 C2

SLIDE 18

October 19, 2007 University of Texas at San Antonio

RMSE = 0.27 pixel RMSE = 0.27 pixel

Before After

SLIDE 19

October 19, 2007 University of Texas at San Antonio

MSS w ith TM MSS w ith TM

SLIDE 20

October 19, 2007 University of Texas at San Antonio

1974 MSS and 2002 ETM+ (0.24 pixel) 1974 MSS and 2002 ETM+ (0.24 pixel) Alaskan ACP (70.46 Alaskan ACP (70.46 oN, 155.25 N, 155.25 oW) W)

SLIDE 21

October 19, 2007 University of Texas at San Antonio

Multi-Decadal Lake Change Record Multi-Decadal Lake Change Record

2000000 4000000 6000000 8000000 10000000 2000000 4000000 6000000 8000000 10000000

1973 Lake Area (m2) 1997 Lake Area (m2)

Stable lake Expanding lake Disappeared lake Shrinking lake

SLIDE 22

October 19, 2007 University of Texas at San Antonio

Expected Results from Pan-Arctic Studies Expected Results from Pan-Arctic Studies Systematic inventory of high-latitude lakes; Metrics on lake dynamics. Science questions and Hypotheses:

“How have northern lakes and wetlands responded to rising Arctic temperatures, and what does their future hold with respect to continued warming in the region?” High-latitude lakes are in a disequilibrium state since the 1970’s. Lake changes are significantly influenced by other factors besides climate, such as permafrost state. The ultimate “endgame” for a hotter Arctic is a shift from above-ground to below-ground storage of water.

SLIDE 23

October 19, 2007 University of Texas at San Antonio

Global Lake Distribution Global Lake Distribution

Largest group of lakes:

high-latitudes;

Second largest:

27 -- 42oN Where are they?

Source: GLWD (Lehner and Doll, 2004)

SLIDE 24

October 19, 2007 University of Texas at San Antonio

“Roof of the World”; Lake density: 2.5%; Warming: 0.16oC per decade; Little anthropogenic impact; Challenging environment for fieldwork.

Tibetan Plateau Tibetan Plateau

SLIDE 25

October 19, 2007 University of Texas at San Antonio

Science Questions Science Questions

How do present-day lake areas compare with maximum lake extents during the GLP period in the late Pleistocene, as evidenced by paleo-shoreline data? How have areas and distributions changed

ver the past 30 years, an interval of

pronounced warming in the region? What are the driving mechanisms underlying the observed lake changes?

SLIDE 26

October 19, 2007 University of Texas at San Antonio

Remote Sensing of Remote Sensing of Paleo Lake Changes Paleo Lake Changes

Greatest Lake Period (GLP): between ~40 and 25 ka BP; Shrunk greatly since then; "How much have the Tibetan lakes shrunk since the late Pleistocene? " Integrated RS/GIS approach.

SLIDE 27

October 19, 2007 University of Texas at San Antonio

Background Background

SLIDE 28

October 19, 2007 University of Texas at San Antonio

Dagze Dagze Co: A Typical Tibetan Lake

: A Typical Tibetan Lake

paleo shores paleo shores

ffspring lakes

lacustrine sediments

243 km2 4466 m a.m.s.l.

SLIDE 29

October 19, 2007 University of Texas at San Antonio

Interactive Paleo Lake Mapping Environment Interactive Paleo Lake Mapping Environment

SLIDE 30

October 19, 2007 University of Texas at San Antonio

Recovered paleo lake extent matching lake features

Paleo water level: ~4523 m with a variation of 3 m; Paleo lake extent: ~846 km2; Water loss: ~ 30.4 km3 water; Paleo lake broke into modern Dagze Co and 30+ lakes.

SLIDE 31

October 19, 2007 University of Texas at San Antonio

Paleo Lake Recovery Across the Plateau Paleo Lake Recovery Across the Plateau

653 contemporary lakes evolved from 173 paleo mega lakes. Total area shrinkage and water loss are estimated at 42,109 km2 and 2,936 km3.

94°E 94°E 92°E 92°E 90°E 90°E 88°E 88°E 86°E 86°E 84°E 84°E 82°E 82°E 80°E 80°E 78°E 36°N 36°N 34°N 34°N 32°N 32°N 30°N 30°N 28°N 28°N

Legend

Modern Lakes Paleo Lakes Tibetan Plateau

SLIDE 32

October 19, 2007 University of Texas at San Antonio

Spatial Pattern of Paleo Lake Change Spatial Pattern of Paleo Lake Change

Zone 1: minor water-level drop (<20 meters). Zone 2: the moderate zone, with 20-60 meter water level drop. Zone 3: greatest water-level drop, up to 285 meters.

94°E 94°E 92°E 92°E 90°E 90°E 88°E 88°E 86°E 86°E 84°E 84°E 82°E 82°E 80°E 80°E 78°E 36°N 36°N 34°N 34°N 32°N 32°N 30°N 30°N 28°N 28°N

Legend

3 - 7 (m) 8 - 10 11 - 15 16 - 20 21 - 30 31 - 40 41 - 50 51 - 60 61 - 70 71 - 80 81 - 100 101 - 130 131 - 200 201 - 300 Tibetan Plateau

SLIDE 33

October 19, 2007 University of Texas at San Antonio

Recent Dynamics Recent Dynamics

f Daw a
f Daw a Co

Co

(a)

(a) ETM image of 10/28/2000; (b) Lake change between 11/15/1976 MSS and 11/10/1990 TM image; (c) Lake change between 10/10/1990 TM and 10/28/2000 ETM+ image.

shrinking expanding (b) further shrinking

further expanding

(c)

SLIDE 34

October 19, 2007 University of Texas at San Antonio

Challenges in Global Lake Dynamics Challenges in Global Lake Dynamics Adequate lake change detection:

Precise image co-registration at sub-pixel accuracy; Accurate lake identification;

Automation:

Accurate lake identification and sub-pixel accuracy co-registration;

Satellite image acquisition; Addressing seasonal variation; Understanding the mechanism of lake changes.

SLIDE 35

October 19, 2007 University of Texas at San Antonio

Critical Techniques Critical Techniques

Algorithms have been tested in Arctic and Tibetan Plateau:

Image co-registration; Lake identification; Change detection.

Automation!

SLIDE 36

October 19, 2007 University of Texas at San Antonio

Challenge to satellite image acquisition Challenge to satellite image acquisition

Suitable 1970’s Landsat coverage is not comprehensive.

Possible Solutions Possible Solutions Making a large budget for projects; Calling for institutional attention; Coordinating among researchers; Encouraging data sharing and trading.

SLIDE 37

October 19, 2007 University of Texas at San Antonio

Challenge to addressing seasonal variation Challenge to addressing seasonal variation

Seasonal variation vs. Long-term changes;

Possible Solutions Possible Solutions Narrow the lake mapping episode to the best season; Avoid the snow melting and flood periods; Leave blanks rather than include images in unwanted seasons; Use overlap area of neighboring scenes to quantify seasonal variations.

SLIDE 38

October 19, 2007 University of Texas at San Antonio

Challenge to understanding the mechanism Challenge to understanding the mechanism behind lake changes behind lake changes

Involved various factors.

Possible Solutions Possible Solutions

Collecting various data sources:

Topographic data; Environmental data; Hydrological data; Climatological data;

Establishing GIS database; Using comprehensive GIS analysis;

SLIDE 39

October 19, 2007 University of Texas at San Antonio

(Mac-Donald et al.,2006) (Brown et al.,1997) GTOPO30 <300 m (Ray and Adams, 2001)

GIS-Based Mechanism Analysis GIS-Based Mechanism Analysis

SLIDE 40

October 19, 2007 University of Texas at San Antonio

Conclusions Conclusions

Remote sensing of lake dynamics:

On-going efforts; Global-scale;

A lot of work! Acknowledgement:

NASA THP; NSF Arctic Science Program; NASA NIP.

SLIDE 41

October 19, 2007 University of Texas at San Antonio