Published online 13 May 2005
Published in Agron J 97:872-878 (2005)
DOI: 10.2134/agronj2004.0162
© 2005 American Society of Agronomy
677 S. Segoe Rd., Madison, WI 53711 USA
Remote Sensing
Predicting Rice Yield Using Canopy Reflectance Measured at Booting Stage
Kuo-Wei Changa,
Yuan Shenb,* and
Jeng-Chung Loc
a No. 70-11, Beishiliao Liau, Beishiliao Village, Madou Town, Tainan County, 721, Aletheia Univ., Taiwan, ROC
b Dep. of Soil and Environmental Sciences, National Chung-Hsing Univ., Taichung, 402, Taiwan, ROC
c Dep. of Agronomy, Chiayi Station, TARI. Chiayi, Taiwan, ROC
* Corresponding author (yshen{at}nchu.edu.tw)
Received for publication June 16, 2004.
 |
ABSTRACT
|
|---|
Abilities to estimate rice (Oryza sativa L.) yields within fields from remote sensing images is not only fundamental to applications of precision agriculture, but can also be very useful to food provisions management. Major objectives of this study were to identify spectral characteristics associated with rice yield and to establish their quantitative relationships. Field experiments were conducted at Shi-Ko experimental farm of TARI's Chiayi Station during 1999–2001. Rice cultivar Tainung 67, the major cultivar grown in Taiwan, was used in the study. Various levels of rice yield were obtained via N application treatments. Canopy reflectance spectra were measured during entire growth period, and dynamic changes of characteristic spectrum were analyzed. Relationships among rice yields and characteristic spectrum were studied to establish yield estimation models suitable for remote sensing purposes. Spectrum analysis indicated that the changes of canopy reflectance spectrum were least during booting stages. Therefore, the canopy reflectance spectra during this period were selected for model development. Two multiple regression models, constituting of band ratios (NIR/RED and NIR/GRN), were then constructed to estimate rice yields for first and second crops separately. Results of the validation experiments indicated that the derived regression equations successfully predicted rice yield using canopy reflectance measured at booting stage unless other severe stresses occurred afterward.
Abbreviations: GNDVI, green normalized difference vegetation index • GRN, green band • MSPR, mean squared prediction error • NDVI, normalized difference vegetation index • NIR, near-infrared reflectance • RED, red reflectance • TARI, Taiwan Agriculture Research Institute
|
INTRODUCTION
|
|---|
PRECISION AGRICULTURE is a recently developed crop management technique, which requires understanding of spatial variability of crop yields and the causes. Through this new technique, agricultural nonpoint pollution can be reduced through efficient farm management practices by applying fertilizers and pesticides at variable rates. Sustainable agriculture can thus be obtained by not only raising production profits but also protecting the environment (Blackmore et al., 1995; Castelnuovo, 1995; Pierce and Sadler, 1997).
Yield maps are the basis of making precision management decisions. Through accumulated yield maps during past seasons, maps for field management can be produced. Regions always having higher or lower yields can be easily delineated, which can be very useful for diagnosing the causes responsible for low yield. Proper management strategies can then be applied. Where available, remotely sensed images showing spatial and spectral variations resulting from soil and crop characteristics are important sources of data for making yield maps (NRC, 1997). A remote sensed yield map would not be affected by the inaccuracies (problems connected with grain flow dynamics and accurate logging of geographic position) associated with combine yield monitors, as suggested by Lark et al. (1997) and Arslan and Colvin (1999). However, difficulty is a lack of valid regression models to convert imagery spectral information to a yield map.
A ratio vegetation index, dividing near-infrared reflectance (NIR) by red reflectance (RED), was one of the early remote sensing techniques to identify the vegetation contribution (Jordan, 1969). The basis of this relationship is the strong absorption (low reflectance) of red light by chlorophyll and low absorption (high reflectance) in the NIR by green leaves (Avery and Berlin, 1992). Normalized difference vegetation index (NDVI), where NDVI = (NIR – RED)/(NIR + RED), was also proposed to estimate green biomass. The basis of NDVI and green biomass appears to be related to the amount of photosynthetically active radiation absorbed by the canopy. The NDVI relates the reflectance in the red region and NIR region to vegetation variables such as leaf area index, canopy cover, and the concentration of total chlorophyll (Tucker, 1979; Sellers, 1985, 1987), which has in turn been associated with crop yield (Wiegand et al., 1994; Aparicio et al., 2000). However, when chlorophyll content, vegetation fraction, or leaf area index reaches moderate to high values, NDVI is less sensitive to these biophysical parameters (Buschmann and Nagel, 1993; Gitelson and Merzlyak, 1994; Aparicio et al., 2000). Gitelson et al. (1996) indicated that, under these conditions, green band (GRN) is more sensitive than the red band and have proposed using a green normalized difference vegetation index (GNDVI), where GNDVI = (NIR – GRN)/(NIR + GRN) for assessing biomass variation. A considerable amount of research with remote sensing of corn (Zea mays L.) canopies (Blackmer et al., 1994; Schepers et al., 1992, 1996) has shown that the green band (in combination with the NIR band) is more highly associated with the variability in leaf chlorophyll, N content, and grain yield than the red band.
Rice is the major staple food for Asian countries. Remote sensing techniques have the potential to provide information on agricultural crops quantitatively, instantaneously, and nondestructively over large areas. Abilities to estimate rice yields within fields from remote sensing images is not only fundamental to applications of precision agriculture, but can also be very useful to governmental administrators for food provisions management. Though many researchers have been devoted to rice planting hectarage estimation (Leblon et al., 1991; Prince, 1991; Bouman, 1992; Wiegand et al., 1992; Field et al., 1995; Clevers and van Leeuwen, 1996; Bach, 1998; Moulin et al., 1998; Reynolds et al., 2000; Serrano et al., 2000), few studies have been conducted attempting to relate canopy reflectance spectra measurements to grain yields. Basic studies regarding the timing of reflectance measurements and regression models for rice yield prediction/estimation are still very lacking. Our objectives were thus to collect the reflectance spectrum of rice canopies, to identify spectral characteristics associated with rice yield, and to establish their quantitative relationships.
|
MATERIALS AND METHODS
|
|---|
The study was conducted at Shi-Ko experimental farm of Taiwan Agriculture Research Institute (TARI) Chiayi Station (23°35'4'' N lat, 120°24' E long) during 1999–2001 growing seasons. The soil at this site is a silt clay loam (mixed hyperthermic Haplaquepts). Paddy rice cultivar Tainung 67, the major cultivar grown in Taiwan, was used in the study. Different N levels were used in each year: 0, 90, and 180 kg N ha–1 for 1999; 0, 30, 60, 90, and 180 kg N ha–1 for 2000; and 0, 45, 90, and 180 kg N ha–1 for 2001, to affect rice yield. Field plots were shifted every year to other well-fertilized production fields to avoid any residue fertilizer effect from the previous year. The plots in each year were all arranged in a randomized complete block design with three replications. Individual plot dimensions were at least 10 by 10 m.
The rice was grown under a conventional two-season cropping system. Three-leaf old rice seedlings were transplanted in early February for the first season crop and in early July for the second season crop. Transplanting density was 0.15 by 0.25 m with three plants per hill. Other than N, all plots were fertilized with 150 kg P2O5 ha–1 and 150 kg K2O ha–1 at the time of transplanting as the basal dose. The N fertilizer was applied as three quotas, were applied as basal at transplanting, and top-dressed at active tillering and panicle initiation stages, respectively. In-season weed and pest controls were practiced according to regional recommendations.
Canopy reflectance spectra were measured every 1 to 2 wk, depending on growth stages, during the entire growth period of each cropping season using a portable spectroradiometer (LI-1800, LICOR) with remote cosine receptor (LI-1800-02, LICOR) attached to an 1.5-m extension arm. The arm was held 1 m above the canopy. At this height, a target area of 1-m radius occupied 80% of the view. The man holding the extension arm always wore dark clothes and standing sideways to reduce measurement error. All the measurements were made near midday, within 2 h of solar noon. Incident and reflected solar radiations were measured by facing the remote cosine receptor upward and downward, respectively. The measurements were taken over the wavelength range from 400 to 1100 nm at a scanning interval of 10 nm and executed consecutively three times per subplot to reduce the possible effect of changing sky conditions. The reflectance of canopies were then calculated from the mean of three repetitions.
Incremental values of spectral reflectance were averaged within 0.50 to 0.59, 0.61 to 0.68, and 0.79 to 0.89 µm to give, respectively, values of green (GRN), red (RED), and near-infrared (NIR) bands of reflectance. Two normalized difference vegetation indices (NDVI and GNDVI) and two ratio vegetation indices (NIR/GRN and NIR/RED) were then calculated. The dynamic changes of these four vegetation indices during growing season were analyzed for each treatment at different year and crop season.
At maturity, grain yield was estimated by hand harvesting three 1-m areas in each. After threshing, fully filled grains sieved by wind-selection method were sun dried for 1 wk. After drying, grain yield per subplot was weighted and adjusted to a constant moisture basis of 13.5% water. Analysis of variance on yields and regression analysis of relationships between yields and vegetation indices were evaluated using the General Linear Models procedure (SAS, Ver. 8; SAS Inst., 1999). Correlation matrices of observed yield, band ratios, vegetation indices, and spectra values were analyzed via ANOVA with a mixed model (Statistica Ver. 6; StatSoft, 2001).
Independent data of canopy reflectance at booting stage and grain yield from new experiments, regarding N rate and water deficit, conducted also at Shi-Ko experimental farm on the two crop growing seasons of 2002, were used to check the predictive ability of the developed models. The experimental design of the N experiments, with levels of 0, 45, 90, and 180 kg N ha–1, were similar to those described above. The water deficit experiments consisted of two treatments, three replications per treatment, of withholding irrigation water from active tillering stage to maturity or from panicle initiation stage to maturity, respectively. Field management practices of the water deficit experiments followed those of 180 kg N ha–1 N experiment but were only rainfed after imposing the water stress. Relative error and mean squared prediction error (MSPR) was then computed to measure the actual predictive capability of the models (Neter et al., 1999, p. 54–56).
|
RESULTS AND DISCUSSION
|
|---|
Typical temporal changes of rice canopy reflectance spectra during first crop season were shown in Fig. 1
. Three weeks after transplanting, rice seedlings were just recovering from damages induced by transplanting. At this time, percent ground cover was <15%. Values of reflectance in the visible region were only slightly lower than those in the near-infrared region because underlying water and soil contributed most to the measured canopy reflectance spectrum. As rice plants grew, contribution from the plants gradually increased. At active tillering stage, 6 and 9 wk after transplanting, tiller number and leaf area increased much more rapidly at a rate of about 1.3 tillers d–1. Accordingly, reflectance in near-infrared region increased rapidly as a result of increased light scattering by leaves and stems. However, the reflectance in the visible region decreased due to absorption by pigments, chlorophyll in particular. About 12 wk after transplanting, reflectance in the near-infrared region reached the highest value of the season while reflectance in the visible portion reached the lowest value. At later stages, yellowing and wilting of rice plants gradually appeared. Therefore, reflectance in the visible region increased as a result of decreasing chlorophyll concentration but reflectance in the near infrared region decreased due to wilting, the exposing of soil background. The reflectance data measured in the second crop season also showed similar changes (data not shown).
As indicated above, the visible reflectance of rice canopy decreased and near-infrared reflectance increased as plants grew. During senescence, the reverse phenomenon occurred. Therefore, curves of ratio indices (NIR/RED, NIR/GRN) and vegetation indices (NDVI, GNDVI), as a function of crop development and N treatments, increased rapidly at vegetative growth stage, reached a short plateau in between panicle formation (
75 DAT) and heading stage (90 DAT), and then gradually declined while maturing (Fig. 2)
. Higher rates of applied N fertilizer would increase not only the green biomass but also the total chlorophyll content of leaves (Lee et al., 2002). The reflectance in the visible region (green and red bands) was negatively correlated with leaf chlorophyll concentration while the reflectance of the near infrared band was positively correlated with leaf area (Guyot, 1990).
View larger version (34K):
[in this window]
[in a new window]
|
Fig. 2. Temporal patterns of (A) NDVI, (B) GNDVI, (C) NIR/RED, and (D) NIR/GRN as a function of N rate, measured in first crop season 2000.
|
|
The main effects of N rate and the interaction between N rate and year were statistically significant for yield (Table 1). Yields were also significantly different among the years and crop seasons. These results indicated that rice yield variability was not only affected by the amount of applied N fertilizers but also by the differences in climatic conditions between the years and crop seasons. The correlation matrix of observed yield, band ratios, vegetation indices, and spectra values is shown in Table 2. The correlation coefficients of observed yield with band ratios and vegetation indices were higher than those of spectra values, which indicated that band ratios and vegetation indices could serve as independent variables for developing grain yield prediction models. However, fundamental statistics of observed yield, band ratios, vegetation indices, and spectra values at the booting stage, as shown in Table 3, indicated that the coefficient of variation of band ratios were larger than those of vegetation indices. This meant that the fitting range of band ratios were wider than those of vegetation indices. Therefore, by calculating band ratios, curves of NIR/RED and NIR/GRN enlarged the effects of N application rate and gave the greatest separation among treatments. Curves of GNDVI and NDVI did not show a good separation among treatments, implying that these two indices were less preferable to serve as independent variables for developing the required grain yield estimation model. Changes of NIR/RED and NIR/GRN curves during the period from panicle formation to heading were few compared with other periods. Therefore, we considered that booting stage, between the panicle formation and heading stages, might be the best period for developing a yield prediction model using canopy reflectance measurements. This implies the boot stage cover 75% of the growth period. It only lasts approximately 7 to 10 d. Rice plants at this stage should have experienced most of the growth limiting factors.
View this table:
[in this window]
[in a new window]
|
Table 1. Analysis of variance for rice yield at three N application rates (0, 90, and 180 kg N ha–1) and two crop seasons in 3 yr (1999–2001).
|
|
View this table:
[in this window]
[in a new window]
|
Table 3. Fundamental statistics of observed yield, band ratios, vegetation indices, and spectrum values at booting stage of first crop of 2000.
|
|
Significant linear relations existed between grain yield and band ratio of NIR/RED or NIR/GRN at booting stages for two growing seasons in each year (Table 4). But, the intercepts and slopes changed across different years and growing seasons, which indicated that a robust simple linear regression equation to predict grain yield by band ratio of NIR/RED or NIR/GRN at booting stage alone was not available. For example, the grain yields of 90 kg N ha–1 treatment of first growing season on 1999 and 60 kg N ha–1 treatment of first growing season on 2000 were both about 8.4 t ha–1, but the corresponding band ratio of NIR/RED or NIR/GRN were quite different (Table 5). In another example, the grain yields of 90 kg N ha–1 treatment of second growing season on 1999 and 60 kg N ha–1 treatment of first growing season on 2000 were quite different, but the corresponding band ratio of NIR/RED or NIR/GRN were very similar.
View this table:
[in this window]
[in a new window]
|
Table 4. Intercept, slope, and coefficient of determination value for the linear regression of grain yields and reflectance indices at booting stage of each individual year and season.
|
|
The pattern of association among the two indices and grain yields indicated that both indices increased with yield but values of NIR/RED had a much larger variation across years and seasons than those of NIR/GRN. It is thus speculated that canopy architecture and/or leaf pigment composition, which in turn changed canopy reflectance spectra, may have altered due to changing climatic conditions. Growth period of first crop and second crop of rice in Taiwan was from February to June and from July to September, respectively. Averaged grain yields of the second crop season rice were generally about 20% lower than first crop season due to different climatic conditions.
Multiple regression was performed, separated into first and second crop season, to estimate grain yield using the ratio indices, NIR/RED and NIR/GRN, at booting stage (Table 6). Figure 3 shows the prediction of grain yields by the developed models. Statistical analysis indicated that there was no significant difference between the mean observed yield and the mean yield estimated for first season crop (t = –0.000047 with 11 df, 
= 0.05) or second season crop (t = –0.2724 with 10 df, = 0.05). Additional statistical analysis (Neter et al., 1999, p. 54–56) indicated that the slope of the linear regression between observed yield and predicted yield for first crop season was not significantly different from 1 (t = –1.0195 with 10 df, = 0.05), whereas the intercept was not significantly different from 0 (t = 7.4101 with 10 df, = 0.05), suggesting that the linear regression through these data was not significantly different from the 1:1 line. Similarly, the slope of the linear regression between observed yield and predicted yield for second crop season was not significantly different from 1 (t = –1.353 with 9 df, = 0.05), whereas the intercept was not significantly different from 0 (t = 3.926 with 9 df, = 0.05), again suggesting that the linear regression through these data was not significantly different from the 1:1 line.
View this table:
[in this window]
[in a new window]
|
Table 6. Regression equations for predicting grain yield by ratio indices, NIR/RED, and NIR/GRN at booting stage.
|
|
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 3. Predicted vs. observed grain yields for (A) first crop season and (B) second crop season of 1999–2001.
|
|
Results of the validation test using independent data sets collected in 2002 were shown in Table 7. Yield overestimation for treatments that were water stressed from panicle initiation was expected because the water stress was imposed only several days before the canopy reflectance measurements and did not have time to produce noticeable spectral changes. Except for this treatment, relative estimation errors were <15% for all other treatments. Mean squared prediction error were 0.40 and 0.22 for first and second crop season, respectively. These values were fairly close to the corresponding mean square error, 0.32 and 0.14, of the regression fit to the model-building data set. Thus, the ratio indices, NIR/RED and NIR/GRN, at booting stage can be considered reliable independent variables for predicting rice yield and the predictive capability of the proposed multiple regression models have been demonstrated.
View this table:
[in this window]
[in a new window]
|
Table 7. Average values of actual rice yield and those predicted by the models for treatments of various nitrogen application rates and water stress levels in the validation study conducted in 2002.
|
|
|
CONCLUSIONS
|
|---|
This study showed that use of indices, NIR/RED and NIR/GRN, derived from canopy reflectance spectra measured at booting stage was highly correlated with grain yield. The derived regression equations also have the potential to predict grain yield for years to come. However, it should be noted that the rice grain yield predicting equations were evaluated only for one variety and under the unique environmental conditions presented in this study. Thus, further validation of the use of these equations under more diverse environmental conditions is needed to warrant its widespread use. If proven to be reliable, actual yield maps can be produced from imagery data with proper reflectance correction. The maps are not only useful for making management decisions, but also for depicting spatial variability in fields before harvest while there is still time to examine the growing crop.
|
ACKNOWLEDGMENTS
|
|---|
Financial support for this research came from projects 89RS-4.1-FAD-41(4), 90AS-4.1.1-CI-C2, and 91AS-5.1.3-CI-C2 of Council of Agriculture, Taiwan, ROC.
|
REFERENCES
|
|---|
- Aparicio, N., D. Villegas, J. Casadesus, J.L. Araus, and C. Royo. 2000. Spectral vegetation indices as nondestructive tools for determining durum wheat yield. Agron. J. 92:83–91.[Abstract/Free Full Text]
- Arslan, S., and T.S. Colvin. 1999. Laboratory performance of a yield monitor. Appl. Eng. Agric. 15:189–195.
- Avery, T.E., and G.L. Berlin. 1992. Fundamentals of remote sensing and airphoto interpretation. 5th ed. Macmillan, New York.
- Bach, H. 1998. Yield estimation of corn based on multitemporal Landsat-TM data as input for an agrometeorological model. Pure Appl. Opt. 7:809–825.
- Blackmer, T.M., J.S. Schepers, and G.E. Varvel. 1994. Light reflectance compared with other nitrogen stress measurements in corn leaves. Agron. J. 86:934–938.[Abstract/Free Full Text]
- Blackmore, B.S., P.N. Wheeler, J. Moris, R.M. Moris, and R.J.A. Jones. 1995. The role of precision farming in sustainable agriculture: A European perspective. p. 777–793. In P.C. Robert et al. (ed.) Site-specific management for agricultural systems. ASA, CSSA, and SSSA, Madison, WI.
- Bouman, B.A.M. 1992. Linking physical remote sensing models with crop growth simulation models, applied for sugar beet. Int. J. Remote Sens. 13:2565–2581.
- Buschmann, C., and E. Nagel. 1993. Spectroscopy and internal optics of leaves as basis for remote sensing of vegetation. Int. J. Remote Sens. 14:711–722.
- Castelnuovo, R. 1995. Environmental concerns driving site-specific management in agriculture. p. 867–880. In P.C. Robert et al. (ed.) Site-specific management for agricultural systems. ASA, CSSA, and SSSA, Madison, WI.
- Clevers, J.G.P.W., and H.J.C. van Leeuwen. 1996. Combined use of optical and microwave emote sensing data for crop growth monitoring. Remote Sens. Environ. 56:42–51.
- Field, C.B., J.T. Randerson, and C.M. Malmstrom. 1995. Global net primary production: Combing ecology and remote sensing. Remote Sens. Environ. 51:74–88.[CrossRef]
- Gitelson, A.A., and M.N. Merzlyak. 1994. Spectral reflectance changes associated with autumn senescene of Aesculus hippocastanum L. and Acer platanoids L. leaves. Spectral features and relation to chlorophyll estimation. J. Plant Physiol. 143:286–292.[ISI]
- Gitelson, A.A., Y.J. Kaufman, and M.N. Merzlyak. 1996. Use of a green channel in remote sensing of global vegetation from EOS-MODIS. Remote Sens. Environ. 58:289–298.[CrossRef]
- Guyot, G. 1990. Optical properties of vegetation canopies. p. 19–43. In M.D. Steven and J.A. Clark (ed.) Applications of remote sensing in agriculture. Butterworths, London.
- Jordan, C.F. 1969. Derivation of leaf area index from quality of light on the forest floor. Ecology 50:663–666.[CrossRef][ISI]
- Lark, R.M., J.V. Stafford, and H.C. Bolam. 1997. Limitations on the spatial resolution of yield mapping for combinable crops. J. Agric. Eng. Res. 66:183–193.
- Leblon, B., M. Guerif, and F. Baret. 1991. The use of remotely sensed data in estimation of PAR use efficiency and biomass production of flooded rice. Remote Sens. Environ. 38:147–158.
- Lee, Y.J., C.M. Yang, and A.H. Chang. 2002. Changes of nitrogen and chlorophyll contents and reflectance spectral characteristics to the application of nitrogen fertilizer in rice plants. J. Agric. Res. China 51:1–14.
- Moulin, S., A. Bondeau, and R. Delecolle. 1998. Combinging agricultural crop models and satellite observations: From field to regional scales. Int. J. Remote Sens. 19:1021–1036.
- National Research Council (NRC). 1997. Precision agriculture in the 21st century: Geospatial and information technologies in crop management. Rep. 59-0700-4-139. NRC, Washington, DC.
- Neter, J., M.H. Kutner, C.J. Nachtsheim, and W. Wasserman. 1999. Applied linear statistical models. McGraw-Hill, New York.
- Pierce, F.J., and E.J. Sadler. (ed.). 1997. The state of site specific management for agriculture. ASA, CSSA, and SSSA, Madison, WI.
- Prince, S.D. 1991. A model of regional primary production for use with coase-resolution satellite data. Int. J. Remote Sens. 12:1313–1330.
- Reynolds, C.A., M. Yitayew, D.C. Slack, C.F. Hutchinson, A. Huete, and M.S. Petersen. 2000. Estimating crop yields and production by integrating the FAO crop specific water balance model with real-time satellite data and ground-based ancillary data. Int. J. Remote Sens. 21:3487–3508.
- SAS Institute. 1999. SAS/STAT user's guide. Version 8. SAS Inst., Cary, NC.
- Schepers, J.S., T.M. Blackmer, W.W. Wihlelm, and M. Resende. 1996. Transmittance and reflectance measurements of corn leaves from plants differing in nitrogen and water supply. J. Plant Physiol. 148:523–529.[ISI]
- Schepers, J.S., D.D. Francis, M.F. Vigil, and F.E. Below. 1992. Comparison of corn leaf nitrogen concentration and chlorophyll meter readings. Commun. Soil Sci. Plant Anal. 23:2173–2187.
- Sellers, P.J. 1985. Canopy reflectance, photosynthesis, and transpiration. Int. J. Remote Sens. 6:1335–1372.
- Sellers, P.J. 1987. Canopy reflectance, photosynthesis, and transpiration: II. The role of biophysics in the linearity of their interdependence. Remote Sens. Environ. 21:143–183.
- Serrano, L., I. Fillela, and J. Penuelas. 2000. Remote sensing of biomass and yield of winter wheat under different nitrogen supplies. Crop Sci. 40:723–731.[Abstract/Free Full Text]
- StatSoft. 2001. STATISTICA (data analysis software system). Version 6. StatSoft, Tulsa, OK.
- Tucker, C.J. 1979. Red and photographic infrared linear combinations form monitoring vegetation. Remote Sens. Environ. 8:127–150.[CrossRef]
- Wiegand, C.L., S.J. Maas, J.K. Aase, J.L. Hatfield, Jr., P.J. Pinter, R.D. Jackson, E.T. Kanemasu, and R.L. Lapitan. 1992. Multisite analyses of spectral–biophysical data for wheat. Remote Sens. Environ. 42:1–21.
- Wiegand, C.L., D.E. Richardson, D.E. Escobar, and J.H. Everitt. 1994. Photographic and video graphic observations for determining and mapping the response of cotton to soil salinity. Remote Sens. Environ. 49:212–223.[CrossRef]
TOP
ABSTRACT
INTRODUCTION
