Estimating Wheat Equivalent Water Thickness Using Landsat TM/ETM+ Data PDF

Preview

Summary

ESTIMATING WHEAT EQUIVALENT WATER THICKNESS USING LANDSAT TM/ETM+ DATA Abduwasit Ghulam a,b,c,*, Tim Kusky a, Qiming Qinb, Zhao-Liang Li c,e, Alimujiang Kasimud a Center for Environmental Sciences, Saint Louis University, St. Louis, MO 63103, USA b Institute of Remote Sensing and GIS, Peking Univers...

Description

Accelerat ing t he world's research.

Estimating Wheat Equivalent Water Thickness Using Landsat TM/ETM+ Data Timothy Kusky IGARSS 2008 - 2008 IEEE International Geoscience and Remote Sensing Symposium

Cite this paper

Downloaded from Academia.edu 

Get the citation in MLA, APA, or Chicago styles

Related papers

Download a PDF Pack of t he best relat ed papers 

APPLICAT ION OF SUPERVISED ENHANCEMENT T ECHNIQUE FOR CROP LAND MAPPING FROM … Sapt arshi Mondal

Comparat ive analysis of different uni- and mult i-variat e met hods for est imat ion of veget at ion wat er … Clement AT ZBERGER A global review of remot e sensing of live fuel moist ure cont ent for fire danger assessment : Moving t o… Philip Zylst ra, Emilio Chuvieco, Mart a Yebra

ESTIMATING WHEAT EQUIVALENT WATER THICKNESS USING LANDSAT TM/ETM+ DATA Abduwasit Ghulam a,b,c,*, Tim Kusky a, Qiming Qinb, Zhao-Liang Li c,e, Alimujiang Kasimud a

Center for Environmental Sciences, Saint Louis University, St. Louis, MO 63103, USA b Institute of Remote Sensing and GIS, Peking University 100871, Beijing, China c Laboratoire des Sciences de l’Image, de l’Informatique et de la Télédétection, LSIIT (UMR7005), 67400, Illkirch, France d Center for Environmental Remote Sensing (CEReS), Chiba University, Chiba 263-8522, Japan e Institute of Geographic Sciences and Natural Resources Research, Beijing 100101, China *Presenting author: Email: [email protected] ; TeL: +1-314-977-7062, Fax: +1-314-977-3568

ABSTRACT Atmospheric corrected Landsat Enhanced Thematic Mapper Plus (ETM+) near-infrared (NIR) and shortwave infrared (SWIR) band reflectances are used to develop a new index to monitor vegetation water content (VWC) in terms of equivalent water thickness (EWT, cm). This paper outlines the first part of a research program to investigate the potential and physical basis of wavelengths in the optical domain to assess the VWC. Then, a method called vegetation water content index (VWCI) were developed using SWIR, and NIR wavelengths of ETM+ data. The relationship between the EWT at canopy level is explored through linking leaf reflectance data obtained from PROSPECT with canopy reflectance from SailH and in-situ measurements. Significant correlations are found between canopy EWT and the developed index for both modeled and ground measured data. Index Terms— Leaf water content, vegetation water content index (VWCI), vegetation water content estimation 1. INTRODUCTION Equivalent water thickness (EWT, cm2 or cm) is an important parameter describing canopy bio-chemical characters and water status. EWT is defined as the ratio between the quantity of leaf water and leaf area, calculated as (FW-DW)/A in field, where FW and DW stands for the leaf fresh weight and dry weight (in grams), respectively, A is a sample leaf area (cm2). Spectral vegetation indices such as normalized difference vegetation index (NDVI) and relative greenness index (RGI) derived from visible (VIS) and near infrared (NIR) wavelengths have been used to estimate vegetation

978-1-4244-2808-3/08/$25.00 ©2008 IEEE

water content for several decades [1]-[4]. However, the chance of success is limited with these indices due to it represents chlorophyll rather than water content since the wavelengths used mainly located in the strong chlorophyll absorption region and high reflectance plateau of leaf reflectance spectra. Therefore, using of longer wavelengths in infrared range, for example, the shortwave infrared (SWIR) reflectance which is more sensitive to EWT than VIS and NIR, has been suggested by many authors [5]-[7]. Estimation of canopy water content in terms of EWT is based on the relationships between water sensitive indices and EWT. Methods of EWT estimation published in remote sensing literature may be summarized as, 1) quasi-physical models including empirical, semi-empirical methods and curve fitting techniques based on the relationship between water indices and EWT [8]-[11], 2) physical models based on radiative transfer [12]-[17]. Empirical, semi-empirical methods are simpler to formulate and applicable to a large temporal dataset with known exactness, however, it is difficult to generalize for regional and global scale where the vegetation biophysical properties are different. Radiative transfer models can provide physical estimation of VWC with its generalizing power and rather accuracy, but, they are complex and require more input parameters that often are expensive to obtain, especially when the huge amount of data need to be considered, it is too time consuming. The objective of this paper is to further explore the potential of NIR and SWIR wavelengths to develop an index for estimating canopy EWT using Landsat Thematic Mapper (TM) and Enhanced Thematic Mapper Plus (ETM+) data. 2. METHODS The theory behind the development of the index is based on both dry matter and water absorption features of NIR and

II - 785

IGARSS 2008

SWIR reflectance and distribution rules of surface targets in two dimensional reflectance space of actual sensor specific NIR, SWIR wavelengths. TM/ETM+ has one band in NIR region located in 0.76-0.90μm (band4) and two bands in SWIR domain ranging from 1.55-1.75 μm (band5) and 2.09-2.35μm (band7). Elvidge & Lyon (1985) [18] showed that there were significant correlations between EWT and reflectance of both band5 and band7 but the strongest with band5. Therefore, reflectance of ETM+ band4 and band5 was used to formulate our new index. NIR–SWIR reflectance space from time series TM/ETM+ images manifested itself in an approximate trapezoidal shape. The validation of the shape can be found in [19]. As shown on Fig.1, there is a base line similar with soil line in NIR–Red scatter plot, which represents the moisture status of the bare surface. Vertex n (n=A, B, E, F, C, D,) corresponds full cover with rich canopy EWT, full cover with low EWT, partial cover with higher EWT, partial cover with lower EWT, saturated bare soil and dry bare soil, respectively. In the figure below, CD represents the direction of moisture severity. The direction orthogonal to NIR–SWI base line represents the change of surface vegetation from bare soil, partial cover to full cover. AC and BD stand for the maximal and minimal water content lines. EWT of a random pixel is quantified by its location parameters in the trapezoid. In other words, the combination of two vectors, parallel and vertical to the NIR–SWI base line, determines EWT. EG represents the parallel vector, in which the longer the EG is, the less canopy EWT exists. The length of EF is related to the perpendicular distance from the NIR–SWI base line, namely, the further the pixel G (NIRG, SWIRG) locates away from the NIR–SWI base line, the shorter the EF is and vice versa. Therefore, the ratio of GF and EF quantifies the amount of EWT as the combination of parallel and vertical vectors against the NIR–SWI base line. We named this ratio the vegetation water content index (VWCI). Assuming M1, M2, I1, I2 refer the slope and interception of maximal and minimal lines AC and BD and, M, I refer the slope and interception of the NIR–SWIR line, VWCI is expressed in following form. VWCI

comparison and validation. Winter wheat was cultivated from September 23 to October 10. The data used in this study include winter wheat leaf biophysical parameters measured on April 17 and May 19 of 2001 and 2004. Leaf biomass was sampled over 0.6 m*0.6 m area in every test site. Sample leaf fresh weight was recorded immediately after cut and leaves were put in water for more than 5 hours in laboratory to allow them regain full turgor and the turgid weight was recorded. Samples were oven-dried at 75ºC for 12 hours or no change in weight was observed by further drying, and then the dry weight was determined. LAI was measured with the CI-203 Portable Laser Area Meter produced by CID, Inc, USA. Chlorophyll (a, b) were extracted in 80% acetone immediately after the samples were carried to laboratory. The absorption of the extracts at wavelengths of 645nm, and 663nm was measured with a Helios spectrophotometer (Thermo Electron Company, Cambridge, United Kingdom). The Chlorophyll concentration (in mg•g-1 of leaf tissue) was calculated with the formula described by Arnon (1949) [20]. Spectral reflectances of wheat canopy, leaf and different bare soil types were collected in the visible and NIR region (400nm– 2500mm) using GER MARK-V portable spectroradiometer simultaneously with the satellite overpass or the leaf biomass sampling made during the experiment. The spectral resolution was 1–5 nm. Each object was measured three times, and an average was taken.

( M 1  M ) u ( NIR  M 2 u SWIR  I 2 ) ( M 1  M 2 ) u ( NIR  M u SWIR)  ( M 1  M ) u I 2  ( M 2  M ) u I1

(1) 3. RESULTS AND DISCUSSION Fig.1 Sketch map of VWCI

3.1. Study site and data collection The study area is located in northeastern suburb of Beijing city of China. Topography of the area is flat plain average 50 m above the sea level. Yearly mean air temperature, evapotranspiration and precipitation of the study area is 11.5 °C, 2050 mm and 622 mm, respectively. Soil types include dark brown soil, loamy clay and padi soil. Each test field was divided into six sections of 30 m*30 m for model

3.2. Relationship between VWCI and EWT In order to understand the responses of VWCI on EWT variations, leaf reflectance was simulated by PROSPECT [13]. PROSPECT requires four input parameters including structure parameter N, chlorophyll content (Ca+b, Pg/cm2), EWT (g/cm2) and dry matter content (DMC, g/cm2). In this study, input variables were randomly selected following a

II - 786

R = 0.6648 n=22

600 400 200

April 17, 2001

0 0

200

400

600

800

1000 2

Ground measured EWT(g/m )

VWCI

0.6

Effect of soil moisture Effect of LAI

0

May 19, 2001

0 400

5

6

500

1000

1500

2000

2500

3000

3500

Fig. 2 Response of vegetation water content on ETM+ derived VWCI. dependent on LAI. Pixel level EWT increases with increasing of LAI, providing that leaf water content does not change over the canopies. The combination of EWT and LAI as EWTcanopy=LAI*EWT corresponds to a quantity of water per unit area in the canopy. Where EWTcanopy is expressed in g/m2, LAI in m2/ m2 and EWT in g/ m2. To compare modeled data with field measured EWTcanopy, VWCI were calculated from 6S atmospheric corrected surface reflectance of April 17 and May 19 TM/ETM+ image of both 2001 and 2004 by EQ. (1). The simulated logarithmic function of VWCI and EWTcanopy was optimized using least square fitting between field measured and modeled data. Then, the logarithmic equation was used to calculate EWTcanopy values over surface measured plots. There were 23 measured plots in April 17 and 13 in May 19, 2001 and 25 plots for April 17 and May 19 of 2004. The performance of the index was shown by the means of the predictive power (R2) and root mean squired error (RMSE). The results revealed that there were strong linear relationships between model estimated and field measured EWTcanopy (Fig. 3). The strongest correlations between VWCI estimated and in-situ EWTcanopy were R2=0.97 (April 17, 2004), however, the least RMSE=37.8 g/m2 was achieved on May 19, 2001.

200

300

4

EWT (kg/m2)

300

200

3

-0.2

R = 0.7365 n=13

100

2

LAI

2

0

1

0.0

RMSE=37.8 g /m2

100

0.4

0.2

500

1000

600

2

500 400

y = 0.1339Ln(x) - 0.4067 R2 = 0.827

VWCI Estimated EWT(g/m )

800

2

0.8

2

2

RMSE=117.9 g /m2

Relationship between VWCI and EWT, soil moisture and LAI

VWCI Estimated EWT(g/m )

1000

VWCI Estimated EWT(g/m )

2

VWCI Estimated EWT(g/m )

lognormal distribution within the maximum and minimum range defined by the PROSPECT. N may be assumed as a constant for a specific crop in a phonological stage. NIR, SIWR reflectances are not at al affected by chlorophyll. At the leaf level, these two parameters were fixed at N=1.5 and Ca+b =50 μg/cm2 based on the mean values derived from field sampling data of wheat. It was allowed that not only the EWT changed in every input data set, but also the dry matter content changed at the same time. Then, canopy reflectance within TM/ETM+ NIR, SWIR range is simulated with SailH [21] incorporating sensor spectral response function. Satellite observation geometry is restricted to nadir views and a spherical leaf distribution angle (LAD) is used. Range of soil moisture variations is determined as 3.31, 5, 12, 17.8, 30%, respectively, by using field measurements and LAI = 0, 1, 2,…, 6. The maximal and minimal water content lines AB and CD can be determined either by time serious satellite observations or radiative transfer simulations. In this study, coefficients for maximal water content line are calculated taking the average value of the data from April and May over the study area while minimal water content line parameters are derived from the image on May 19, 2001 since this time corresponds the stressed period when the limited irrigation treatment was implemented to monitor the drought effects on the crop. Simulation results indicate that there are strong logarithmic relations between VWCI and EWT, R2=0.827. However, soil moisture causes certain disperse on VWCI over sparsely vegetated surfaces. Such a interference decreases with the increasing of LAI and is almost negligible after LAI reaches 3 (Fig.2). Since the studying of soil moisture effect on the developed index is beyond the scope of this paper, an overall trend-fitting equation is implemented as a generic model for the estimation of canopy EWT. The logarithmic functions obtained from PROSPECT-SailH simulated data were then used to provide estimates of the predictive power of the relationships measured as the R2 between field measured and model estimated EWT. Canopy EWT is also

RMSE=105.6 g /m2

800

2

R = 0.9271 n=25

600 400 April 17, 2004

200 0 0

200

400

600

800

1000

RMSE=43.4 g /m2

450

R2 = 0.8576 n=25

300 150

May 19, 2004

0 0

150

300

450

600

2

2

Ground measured EWT(g/m )

Ground measured EWT(g/m )

Ground measured EWT(g/m2)

Fig. 3 Comparison of index values against ground truth data vegetation water content estimation in terms of Equivalent water thickness. The prediction accuracy may be improved 4. CONCLUSIONS if the most sensitive wavelengths as stated in [20], that only Comparison between VWCI estimated (EWTcanopy) and field available with hyperspectral sensors (e.g. Hyperion and measured data indicated that VWCI has a potential in

II - 787

AVIRIS), implemented to calculate the VWCI. Additional work is required to quantify and eliminate the effects of soil moisture on the proposed index. [12] 5. ACKNOWLDEGEMENTS This work was supported by NSFC (40771148). [13] 6. REFERENCES [1]

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

Burgan, R. E., “Use of remotely sensed data for fire danger estimation,” EARSeL Advances in Remote Sensing, 1996, 4 (4), 1-8. Hardy, C.C. and Burgan, R.E., “Evaluation of NDVI for monitoring live moisture in three vegetation types of the western U.S.,” Photogrammetric Engineering and Remote Sensing, 1999, 65, 603–610. Illera, P., Fernandez, A., & Delgado, J. A., “Temporal evolution of the NDVI as an indicator of forest fire danger,” International Journal of Remote Sensing, 1996, 17 (6), 1093-1105. Paltridge, G.W., & Barber, J., “Monitoring grassland dryness and fire potential in Australia with NOAA/AVHRR data,” Remote Sensing of Environment, 1998, 25, 381-394. Datt, B., “Remote sensing of water content in Eucalyptus leaves,” Australian Journal of Botany, 1999, 47, 909-923. Gao, B., “NDWI-a normalized difference water index for remote sensing of vegetation liquid water from space,” Remote Sensing of Environment, 1996, 58, 257-266. Hunt Jr., E. R., Rock, B. N., & Nobel, P. S., “Measurement of leaf relative water content by infrared reflectance,” Remote Sensing of Environment, 1987, 22, 429-435. Chen, D.Y., Huang, J.F., Jackson T.J., “Vegetation water content estimation for corn and soybeans using spectral indices derived from MODIS near- and short-wave infrared bands,” Remote Sensing of Environment, 2005 , 98 (2-3), 225-236. Champagne, C.M., Staenz, K., Bannari, A., McNairn, H., Deguise, J., “Validation of a hyperspectral curve fitting model for the estimation of plant water content of agricultural canopies,” Remote Sensing of Environment, 2003, 87, 148-160. Jackson, T.J., Chen, D., Cosh, M., Li, F., Anderson, M., Walthall, C., Doriaswamy, P., Hunt, E. R., “Vegetation water content mapping using Landsat data derived normalized difference water index for corn and soybeans,” Remote Sensing of Environment, 2004, 92, 475-782. Sims, D. A., & Gamon, J. A., “Estimation of vegetation water content and photosynthetic tissue area from spectral reflectance: A comparison of

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

II - 788

indices based on liquid water and chlorophyll absorption features,” Remote Sensing of Environment, 2002, 84, 526– 537. Fourty and Baret, “Vegetation water and dry matter contents estimation from top-of the atmosphere reflectance data: a simulation study,” Remote Sensing of Environment, 1996, 61:34-45. Jacquemoud S., Baret F., Andrieu B., Danson F.M., Jaggard K., “Extraction of vegetation biophysical parameters by inversion of the PROSPECT+SAIL model on sugar beet canopy reflectance data – Application to TM data,” Remote Sensing of Environment, 1995, 52:163-172. Riãno, D. Vaughan, P., Chuvieco, E., Zarco-Tejada, P. J., and Ustin, S.L., “Estimation of fuel moisture content by inversion of radiative transfer models to simulate equivalent water thickness and dry matter content: analysis at leaf and canopy level,” IEEE Transactions on Geoscience and Remote Sensing, 2005, 43 (4), 819-826. Zarco-Tejada, P. J., Rueda, C.A. and Ustin, S.L., “Water content estimation in vegetation with MODIS reflectance data and model inversion methods,” Remote Sensing of Environment, 2003, 85, 109-124. Ceccato, P., Gobron, N., Flasse, S., Pinty, B., & Tarantola, S., “Designing a spectral index to estimate vegetation water content from remote sensing data: Part 1. Theoretical approach,” Remote Sensing of Environment, 2002, 82: 188-197. Ceccato, P., Flasse, S., & Gregoire, J. M., “Designing a spectral index to estimate vegetation water content from remote sensing data: Part 2. Validation and applications,” Remote Sensing of Environment, 2002b, 82,198– 207. Elvidge, C.D., & Lyon, R.J.P., “Estimation of the vegetation contribution to the 1.65/2.22μm ratio in airborne thematic-mapper imagery of the Virginia Range, Nevada,” Intrnational Journal of Remote Sensing, 1985, 6, 75–88. A. Ghulam, Li Zhao-Liang, Qin Qiming, Tong Qingxi, Wang Jihua, Kasimu Alimujiang and Zhu Lin, “A method for canopy water content estimation for highly vegetated surfaces–shortwave infrared per...