integrating satellite remote sensing and in-situ ... · decision-making kenya & other african...

1

Integrating Satellite Remote Sensing and In-situ Measurements for Monitoring Surface Soil Moisture

John J. Qu, Xianjun Hao, Rui Zhang, and Ray MothaGlobal Environment and Natural Resources Institute (GENRI)

George Mason University, Fairfax, VA 22030, [email protected]

Robert Stefanski Agricultural Meteorology Division, World Meteorological Organization

Geneva, SWITZERLANDOrivaldo Brunini

Agronomic Institute, BRAZILJohan Malherbe

ARC, LNRPretora, SOUTH AFRICA

2014/4/9 2

Outline

Introduction

Soil Moisture Moisture Monitoring with MODIS

Soil Moisture- Remotely Monitoring and Analysis System (SM-RMAS)

Integrated SMOS and MODIS SM data

Africa Soil Moisture Project

Summary and Discussions

2014/4/9 3

Overview of Soil Moisture Monitoring

Soil moisture is defined as the amount of water in soil.

Soil moisture is an important factor for agricultural drought.

Soil moisture is one of important climate change indicators.

Soil moisture is a critical parameter for IPCC and GFCS.

2014/4/9 4

Overview of Soil Moisture Measurements

In-situ ground measurements Point measurements

Restricted to discrete observation

SM measurement standards and guidelines

Satellite remote sensing Provide large scale and long term observation

Poor spatial resolution for Microwave

Lack of physical basis for Optical/IR

Vegetation cover makes it complex

2014/4/9 5

Soil Moisture Monitoring with MODIS

Combining MODIS and ground measurements for soil moisture estimation Wang et al., 2007a. International Journal of Remote Sensing

Using multiple NIR-SWIR channels to estimate soil and vegetation moisture Wang et al., 2007b. International Journal of Remote Sensing Wang and Qu, 2007. GRL

2014/4/9 6

>11 years (2003 ~ present) data from 137 ground stations 10 cm, 20 cm and 40 cm soil moisture at 8 AM on every 6th, 16th, and 26th

Ground Measurements (~100km2/station)

Data:

MODIS NDVI and LST: 1 km resolution

Remote sensing

Surface cover and soil type 1 km resolution

Soil Moisture Estimation Using MODIS and In-situ Measurements

Study Area: Shandong, China

7

Soil Moisture Estimation Using MODIS and Ground Measurements

T*

NDVI* Low soil moisture

High soilmoisture

NDVIo

NDVIs

To Ts

Universal Triangle (Semi-Physical)(Carlson et al. 1994, Chauhan et al. 2003)

)*(

0

)*(

0

jni

i

inj

jji TNDVIa

os

o

TTTT

T

*

os

o

NDVINDVINDVINDVI

NDVI

*

2014/4/9

θ: soil moisture

(1)

(2)

(3)

The vegetation T is always close to air T, the spatial variation in surface T is small (except for emission from underlying bare soil) over a full vegetation. The surface T over bare soil varies depending on θ: from warm to cold when θ increases.

2014/4/9 8


Regression Model

)*(2

0

)*(2

0

ji

i

ij

jji TNDVIa

Soil Moisture Estimation Algorithm

Soil Moisture from Ground Observations

Regression Algorithm

MODIS NDVI, LST

Soil Moisture at MODIS Resolution

θ=F(NDVI,LST)Regression Coefficients

MODIS Soil Moisture Product Flowchart

2014/4/9 9

Soil moisture product at MODIS resolution


2014/4/9 10

The successful application of ‘Universal Triangle’ relation reinforced this theory;

A good agreement between ground observed and MODIS derived soil moisture suggested soil moisture retrieval by combining in-situ and MODIS measurements is feasible;

The soil moisture map at 1 km resolution provides more regional soil moisture details and spatial pattern.


2014/4/9 11

Moisture effect on canopy reflectance

A new drought index (NMDI) for monitoring soil and vegetation moisture

Integration of multi-space sensor, such as SMOS, SMAP and MODIS and in-situ soil moisture measurements

2014/4/9 12

Soil Moisture Estimation Using Multiple Sensor Measurements

SWIR Bands are

0

0.2

0.4

0.6

0.8

1

400 800 1200 1600 2000

Wavelength (nm)

U.S. 1976 Standard Atmosphere

Atm

osph

eric

Tra

nsm

ittan

ce

SWIR Bands

13Credit to Dr. Menghua Wang NOAA/STAR

14

Limitations:

Cannot remove soil effects Very few similar studies focused on soil moisture

NIR-SWIR vegetation water indices:

NDWI, NDII, WI, GVWI…

Soil Moisture Estimation Using Multiple MODIS SRB Measurements

2014/4/9

2014/4/9 15

Moisture effect on canopy reflectance

PROSPECT

SAIL

Leaf Reflectance

Soil Model

Soil Reflectance

Canopy Reflectance

Coupled soil-leaf-canopy model

(Jacquemund 1990)

(Verhoef 1984)

(Lobell 2002)


2014/4/9 16

NMDI (Normalized Multi-band Drought Index) for remotely sensing soil and vegetation moisture

Formation of NMDI

)()(

21301640860

21301640860

nmnmnm

nmnmnm

RRRRRRNMDI

Simulated soil spectra at various θ and canopy spectra at different Cw

θ R1640nm-R2130nm

Cw (cm)

0

5

10

15

20

25

30

35

500 900 1300 1700 2100

Wavelength [nm]

Spec

tral R

efle

ctan

ce [%

]

0.008

0.016

0.024

0.032

0.04

N=1.3Cab=40μg/cm2

Cm=0.1g/cm2

LAI=2

θ=25%

Cw

R1640nm-R2130nm

θ: soil moisture; Cw: leaf water content

Cw (cm)

0

5

10

15

20

25

30

35

500 900 1300 1700 2100

Wavelength [nm]

Spec

tral R

efle

ctan

ce [%

]

0.008

0.016

0.024

0.032

0.04

N=1.3Cab=40μg/cm2

Cm=0.1g/cm2

LAI=2

θ=25%

Cw

17

NMDI turns to be a vegetation water index at LAI≥2.

NMDI increases almost linearly with Cw, presenting parallel trends for each LAI.

Sensitivity of NMDI to vegetation areas

LAI

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0 0.01 0.02 0.03 0.04 0.05

Cw(cm-1)

NM

DI

0.5123456

R1640nm-R2130nm

2014/4/9

Cw: leaf water content; LAI: leaf area index

2014/4/9 18

0.5

0.6

0.7

0.8

0.9

1

0 0.03 0.06 0.09 0.12 0.15 0.18 0.21Soil Moisture (Vol)

NM

DI

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 0.2 0.4 0.6Soil Moisture (Vol)

NM

DI

Higher NMDI, drier soil. θ <0.1,NMDI = 0.7~1; θ=0.2, NMDI≈ 0.6; θ >0.5, NMDI<0.4. Suggest that soil moisture conditions can be monitored by NMDI with different thresholds.

Validation: MODIS Soil Moisture Products

θ: soil moisture

2014/4/9 19

NMDI is well suited to estimate both soil and vegetation drought.

NMDI demonstrated high performance and discrimination power for fire detection.

The new index will provide a new foundation for monitoring soil moisture for the next generation of MODIS sensor –VIIRS.


SM-RMAS (Soil Moisture- Remotely MonitoringMonitoring and Analysis is a near real time operational system including: (1) satellite remote sensing data processing sub-system, (2) ground measurement database; (3), SM computing sub-system, (4) SM prediction sub-system, and (5) SM analysis and DSS sub-system.

Africa Soil Moisture Project

Agricultural Extension

.

GMU, USDA & UFl Crop Modeling & DSS Tools

for Data Management

Decision-Making

Kenya & Other African Nations

Satellite Remote Sensing Data

Soil & Crop Moisture

Data Products

Decision SupportSystem

User Community

Policy Making

Farm Decisions

Rain GaugeOn Site

Data

IBIMETSNU/NCAM

WAMIS

WC&F Seminars

andTraining

National Drought Policy

Drought/Flood/Heat

Drought/Flood/Heat

Integrated GIS

AgrometData Products

Extension &Training

Satellite Remote Sensing

Decision Support

Soil/Vegetation Moisture Estimation

Drought Assessment

Ground Measurements

Africa SM System Infrastructure

http://wamis.gmu.edu/Africa/

South Africa SM Project Workshop

In-situ Soil Moisture Monitoring Sites in South Africa

Soil Moisture Monitoring in South Africa

Soil Moisture Measurements from SMOS/MODIS

Vegetation Condition Index Temperature Condition Index

Drought Severity Index Evapotranspiration

Soil Moisture Monitoring in Brazil

Soil Moisture Measurements from SMOS/MODIS

2014/4/928Soil Moisture Measurements from SMOS/MODIS

Soil Moisture Monitoring in Mozambique

Soil Moisture Processing System Infrastructure

Satellite & Ground Measurements

Remote Sensing data

SurfaceInformation

Ground Measurements

Data Processing System

Real‐time RSData Processing

Level 2 Data Production

RS DataAnalysis

Inversion Model Cal/Val

Moisture Content Estimation

Data Products

Soil Moisture

Vegetation Moisture

Drought Index

Drought Rating

Application and Decision Support

Drought Assessment

Drought Forecasting

Decision Support

Africa SM Data Dissemination

Public web site Provide GIS based web services to public users and

agencies Dedicated data transfer protocols For institutions and agencies to get data for decision

support and further analysis Mobile applications For public users to view information on mobile devices

2014/4/9 31

Summary and Discussion

Soil moisture estimation by combining the strengths of SMOS, SMAP, MODIS and ground measurements to achieve higher accuracy and spatial resolution.

System (SM-RMAS) is under development. We need to work with our international partners to collect more ground measurements.

We have tested our approach with SMOS and MODIS & in-situ measurements applying it in multiple regions, including China, South Africa, Brazil, and Mozambique.

There is a urgent need to develop standards and guidelines for global soil moisture measurements.

Thank You !

2014/4/9 33

Di Wu, 2014, An Investigation of Agricultural Drought on the United State Corn Belt Using Satellite Remote Sensing and GIS Technology, GMU Ph.D Thesis

Wang, Lingli, 2008, Remote Sensing Techniques for Soil Moisture and Agricultural Drought Monitoring, GMU Ph.D. Thesis.

Dasgupta, Swarvanu, 2007, Remote Sensing Techniques for Vegetation Moisture and Fire Risk Estimation, GMU Ph.D. Thesis.

Hao, Xianjun, 2006, Estimation of Live Fuel Moisture and Soil Moisture Using Satellite Remote Sensing, GMU Ph.D. Thesis.

Soriano, Melissa, 2008, Estimation of Soil Moisture in the Southern United States in 2003 Using Multi-Satellite Remote Sensing Measurements. GMU, Master Thesis

Li, Min 2010, Remote Sensing Techniques for Detecting Vegetation Phenology, GMU Ph.D. Thesis.

Wang, Wanting, 2009, Satellite Remote Sensing of Forest Disturbances Caused by Hurricanes and Wildland Fires. GMU Ph.D. Thesis.

Xie, Yong, 2009, Detection of Smoke and Dust Aerosols Using Multi-sensor Satellite Remote Sensing Measurements. GMU Ph.D. Thesis

Related Publications

2014/4/9 34

Acknowledgments

This project has been funded by WMO and UCAR, NOAA/NWS.

We appreciate the supporting from our GENRI/ESTC partners, WMO, NWS/WR, USDA/FS/SRS, UADA/WAOB, USDA/ARS and international partners, such as ARC, South Africa and Agronomic Institute, Brazil.

integrating satellite remote sensing and in-situ ... · decision-making kenya & other african...

Documents