EG2234: Earth Observation Interactions - Land Dr Mark Cresswell.

EG2234: Earth Observation

Interactions - LandDr Mark Cresswell

Topics

Land surfaces Temperature and radiance Proxy air temperature Albedo NDVI Fires and volcanoes

Land Surfaces

Remote sensing provides a proxy for the nature of land surfaces

We are interested in land cover rather than land use

Different materials (land cover classes) absorb and reflect electromagnetic radiation differently

We can infer land surface conditions from remotely sensed images

Temperature and Radiance

Infrared radiation is the key A region of the electromagnetic spectrum sandwiched

between the red visible and microwave portions of the spectrum

IR radiation is invisible to Humans 3 – 14µm is TIR (0.7 – 1.3µm is NIR, 1.3 – 3µm is MIR, 10.5 - 12.5µm is TIR)

Quantity of IR is related to radiance Radiance intensity is related to temperature

Temperature and Radiance

Detection of infrared energy is performed by the radiometer

Measurement is the result of a current effect due to the direct interaction of photons with electrons present in the detector

Operation of the detector requires that incident photons liberate charge-carrier electrons

Temperature and Radiance

Different materials in nature absorb and emit thermal energy differently

Differences are due to thermal capacity and ability to conduct as well as environmental parameters

Water often has a high thermal capacity compared with bare soil – so both materials are easily distinguished from one another

TIR image

Raw MeteosatRadiance image

Proxy Air Temperature

LST (land surface temperature) may be converted to a proxy air temperature by means of a solar correction algorithm

Knowing the position of the Sun at the time an image is acquired allows the LST to be transformed into a proxy for air temperature (approx 1m above the ground)

WMO Vs Meteosat Temperatures

R2 = 0.8746

-10

-5

0

5

10

15

20

25

30

35

40

0 5 10 15 20 25 30

WMO Temperature

Met

eosa

t T

emp

erat

ure

N = 58

Temperatures in Celsius

95% Confidence Level

WMO Vs Meteosat Temperatures

-5

0

5

10

15

20

25

30

35

40

Time

Tem

pera

ture

(cel

siu

s)

WMO METEOSAT

WMO Vs Corrected Meteosat Temperatures

-5

0

5

10

15

20

25

30

Time

Tem

pera

ture

(ce

lsiu

s)

WMO

CORRECTED

Albedo

Albedo is simply the ratio of incoming radiation to reflected outgoing radiation expressed as %

100in

outA

Albedo

Materials such as clouds and fresh snow have a HIGH albedo (they reflect much of the incident solar radiation back to space)

Forests and soil have a low albedo (they absorb much of the incident radiation)

We can classify materials according to their albedo

Table 1 - Typical Albedo values for natural surfaces

Surface or Object Albedo (%) Fresh snow 75 – 95 Clouds (Thick) 60 – 90 Cloud (Thin) 30 – 50 Ice 30 - 40 Sand 15 – 45 Grassy Field 10 – 30 Water 10 Forest 3 – 10 Earth/Atmosphere 30 Moon 7

VISIBLE

•Albedo•Weather Fcst.

NDVI

Normalised Difference Vegetation Index NDVI provides a good assessment of

photosynthesising vegetation – but caution must be exercised with this type of index as other factors can affect the NDVI other than leaf reflectance: Viewing angle, Soil background, Atmospheric degradation and Leaf orientation

INFRARED REGION

NDVI

Often derived from NOAA-AVHRR satellite system (polar orbiter)

Uses differential reflectance of visible and IR

2 1

2 1

IR R CH CHNDVI

IR R CH CH

CHANNEL SPECTRAL LIMITS REGION

1 0.58 - 0.68µm visible2 0.72 - 1.10µm near infrared3 3.55 - 3.93µm thermal infrared4 10.3 - 11.3µm thermal infrared5 11.5 - 12.5µm thermal infrared

NDVI

NDVI

One of the problems associated with remote sensing vegetation is mixed pixels – a combination of both soil and vegetation

Hutchinson (1982) found that it was difficult to separate soil and vegetation reflectance when vegetation cover is less than 30%.

Solution is Perpendicular Vegetation Index (PVI) (Richardson and Wiegand, 1977)

PVI (perpendicular vegetation index):

R E D R E FLE C TA N C E

NE

AR

IN

FR

AR

ED

RE

FL

EC

TAN

CE

AB

C

W

X Y

soil b righ tness lineB = dark wet soils

C = dry soils

X = pure vegetation pixel

Y = mixed pixel

Fires and Volcanoes

Because of their thermal anomaly, fires and active volcanoes can be identified from space

Acquisition of thermal infrared images allows fires to be detected automatically

Time-series of images can show a trend in temperature beneath and surrounding a volcano that might allow a prediction of eruption to be made

Volcanoes

The moderate resolution imaging spectroradiometer (MODIS) instrument on board the NASA EOS platform, Terra gives global coverage every 1-2 days at 250, 500 and 1000 metre resolutions

Spectral data measured at 4µm and 12µm System examines scenes for high temperature

volcanic thermal anomalies

Volcanoes

Once identified, details such as location, emitted spectral radiance and other parameters are transferred via the internet to the Hawaii Institute of Geophysics and Planetology

Algorithms appear to be robust at detecting both permanent and sporadically active volcanic systems. See Wright et al. 2002

Forest Fires

A fire detection and management system should have the following aims:

• A measure of the geographical limits of the fire-front• An estimate of fire intensity• Monitoring of burnt area to look for latent fires• Mapping of burnt areas to aid restoration

(Barducci et al. 2002)

Borneo Fires (Sep 18th 2005)NASA Earth Observatory, 2005