Trend in Remote Sensing Technology for Forest Measurement & Monitoring Gen Takao CIFOR.

Trend in Remote Sensing Technology for Forest Measurement & Monitoring

Gen TakaoCIFOR

It’s me!big

Trees are …

Measurement is

a hard job

Visibility is

limited

Area isvast

Changes occurslowly & for long

Can be destroyedinstantly

Forest remote sensing

• simultaneity– can observe remote area at once

• periodicity– can observe repeatedly

• recordability– can observe retrospectively

Contents

• How do we measure forests using RS?

• Modern sensors and technique of RS

• Do we actually tell you the truth by RS?

Scales of forest and remote sensing

MODIS 分解能 500m グリッド 50 km

札幌

小樽

羊蹄山 支笏湖

グリッド 5 kmTM/ETM+ 分解能 30m x 1x 2x 4x 8x 16x 32x 64 グリッド 500 mx 128x 256IKONOS PS 分解能 1m x 512x 1024x 2048 グリッド 50 mx 4096

community

1day

10days

(3.3mt) 100days

(2.7yr) 1Kdays

(27.4yr) 10Kdays

1mt

1yr

10yr

1m 10m 100m 1km 10km 100km 1,000kmSpatial

Temporal

individ.trees

smallwatershed

largewatershed

Temporal & spatial scales ofthe sensors

LANDSATMSS/TM/ETM+

NOAAAVHRR

ALOSPRISM

Sensor

Operationlifetime

Observationfrequency

Spatial

resolution

Observation sw

ath

SPOTHRV-P

Terra/AquaMODIS

IKONOS/QuickBird

1 日

10 日

(3.3 月 ) 100 日

(2.7 年 ) 1 千日

(27.4 年 ) 1 万日

1 月

1 年

10 年

1m 10m 100m 1km 10km 100km 1,000km空間

時間

individ.trees

communitysmallwatershed

largewatershed

Temporal & spatial scales ofthe sensors

LANDSATMSS/TM/ETM+

NOAAAVHRR

ALOS PRISMSensor

Operationlifetime

Observationfrequency

Spatial

resolution

Observation sw

ath

SPOT HRV-P

Terra/AquaMODIS

IKONOS/QuickBird

LowMedHigh Sp

atial resolution

Trade-offs to date ofspatial and temporal resolution

• Low spatial resolution ( 100m)≧– Temporal: high globally

• Rapid changes in a wide area, e.g. phenology, fires

• Medium spatial resolution ( 10m)≧– Temporal: not so high globally

• Current condition/temporal changes of vegetation

• High spatial resolution ( ＜ 10m)– Temporal: can be very high at specific targets, but

very low globally• Current condition of an area of interest

Relative size of the spatial resolution to the objects

• Resolution >> Object (L-resolution)– Averaged objects

• Forest canopy

• Resolution << Object (H-resolution)– A piece composing an

object• Individual tree crown

• Resolution ～ Object– Maximum variance

Crown sizes are so variable that often you cannot choose the relative size!

Why do we want to know forest?

• Because we want to manage it for us!

Conserve valuable ecosystems

See beautiful sceneries

Need timbers

Raise profits

Off-limit

Connect pts by walkways

Harvest, re- & afforestation

Slash & burn, golf courses

Aims / incentives Actions

Study the ecophysiology Experiments

Forest management, plan and measurement

A forest now A would-be forest

management

PlanningPlan

ImplementationDo

EvaluationSee

Inventory Monitoringmeasurement

Remote Sensing is Up!