IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 44, NO. 6, JUNE 2006
1545
High-Resolution Change Estimation of Soil Moisture Using L-Band Radiometer and Radar Observations Made During the SMEX02 Experiments Ujjwal Narayan, Venkataraman Lakshmi, Senior Member, IEEE, and Thomas J. Jackson, Fellow, IEEE
Abstract—The soil moisture experiments held during June–July 2002 (SMEX02) at Iowa demonstrated the potential of the L-band radiometer (PALS) in estimation of near surface soil moisture under dense vegetation canopy conditions. The L-band radar was also shown to be sensitive to near surface soil moisture. However, the spatial resolution of a typical satellite L-band radiometer is of the order of tens of kilometers, which is not sufficient to serve the full range of science needs for land surface hydrology and weather modeling applications. Disaggregation schemes for deriving subpixel estimates of soil moisture from radiometer data using higher resolution radar observations may provide the means for making available global soil moisture observations at a much finer scale. This paper presents a simple approach for estimation of change in soil moisture at a higher (radar) spatial resolution by combining L-band copolarized radar backscattering coefficients and L-band radiometric brightness temperatures. Sensitivity of AIRSAR L-band copolarized channels has been demonstrated by comparison with in situ soil moisture measurements as well as PALS brightness temperatures. The change estimation algorithm has been applied to coincident PALS and AIRSAR datasets acquired during the SMEX02 campaign. Using AIRSAR data aggregated to a 100-m resolution, PALS radiometer estimates of soil moisture change at a 400-m resolution have been disaggregated to 100-m resolution. The effect of surface roughness variability on the change estimation algorithm has been explained using integral equation model (IEM) simulations. A simulation experiment using synthetic data has been performed to analyze the performance of the algorithm over a region undergoing gradual wetting and dry down. Index Terms—Hydrology, microwave measurements, moisture change.
I. INTRODUCTION
S
OIL moisture is an important variable in analyzing the dynamics of the atmospheric boundary layer, weather, and climate [1], [2]. Observations of soil moisture are needed at various scales for hydrologic modeling, weather forecasting, climate prediction, flood and drought monitoring, and other water and energy cycle applications [3]–[5]. Active and passive microwave remote sensing provides a unique capability for obtaining frequent observations of soil moisture at global and
Manuscript received January 13, 2005; revised November 27, 2005. This work was supported by NASA Headquarters under the Earth System Science Fellowship Grant NGT5-ESSF/05-0000-0317 awarded to Ujjwal Narayan. U. Narayan and V. Lakshmi are with the Hydroclimatology and Remote Sensing Laboratory, Department of Geologcial Sciences, University of South Carolina, Columbia, SC 29201 USA (e-mail:
[email protected]). T. J. Jackson is with the USDA-ARS Hydrology and Remote Sensing Laboratory, Beltsville, MD 20705 USA. Digital Object Identifier 10.1109/TGRS.2006.871199
regional scales [6]–[8], and the retrieval of soil moisture using ground-based or aircraft mounted radiometers operating at L-band has been demonstrated in several prior studies [9]–[14]. However, currently, operational and proposed L-band satellite radiometers have the problem of moderately coarse spatial resolution which limits their potential applications such as incorporation of soil moisture estimates in agriculture or initializing mesoscale weather models [15]. Radars are capable of much higher spatial resolution than radiometers especially with synthetic aperture processing. However, retrieval of soil moisture using radar backscattering coefficients is difficult due to more complex signal target interaction associated with measured radar backscatter data, which is highly influenced by surface roughness and vegetation canopy structure and water content. Several empirical and semiempirical algorithms for retrieval of soil moisture from radar backscattering coefficients have been developed, but they are valid mostly in the low-vegetation water content conditions [16]–[18]. On the other hand, the retrieval of soil moisture from radiometers is well established and has a better accuracy with limited requirements for ancillary data [3], [19]. Radiometer measurements are less sensitive to uncertainty in measurement and parameterization of surface roughness and vegetation canopy interaction. However, the spatial resolution of radiometer is much lower compared to radar operating in the same band. An optimal soil moisture retrieval algorithm that combines the higher spatial resolution of radar with higher sensitivity of a radiometer might result in improved soil moisture products. Temporal evolution of soil moisture can be potentially monitored through change detection. Change detection methods have been suggested [20] and implemented as a convenient way to determine relative soil moisture or the change in soil moisture [21], [22]. These techniques rely on the assumption that the change in measured brightness temperature or radar backscatter is only due to change in soil moisture at the target. Both brightness temperature and radar backscatter change depend approximately linearly on soil moisture and, hence, sensitivity can be assumed to be independent of soil moisture. However, quantification of sensitivity requires soil moisture measurements, which is difficult in the case of radar in the presence of moderate to high vegetation cover. It may be possible to estimate the radar sensitivity to soil moisture by using radiometer estimated soil moisture measurements if the impact of vegetation on sensitivity and spatial heterogeneity issues can be accounted for. This paper proposes a simple algorithm that uses higher resolution radar observations along with coarser resolution radiometer observations to determine the change in soil moisture
0196-2892/$20.00 © 2006 IEEE
1546
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 44, NO. 6, JUNE 2006
at the spatial resolution of radar operation, without using any in situ soil moisture measurements. The present study simplifies the problem of spatial disaggregation of soil moisture by considering that the spatial variability of bare soil properties (texture, roughness) that influence radar sensitivity to soil moisture is not significant and, hence, the variability of radar signal within the radiometer footprint is due to soil moisture and canopy vegetation water content variability only. The paper is organized into six sections. The next section explains the theoretical basis and assumptions behind the algorithm for spatial disaggregation used in this study. Section III presents the data and the methods that are applied to evaluate the performance of the algorithm presented in the study. Section IV presents the results in terms of comparison of in situ measurements of soil moisture with the disaggregated estimates obtained from the algorithm. Section V describes the design and results of an observation system simulation experiment that was performed to analyze the performance of the algorithm over a longer time period and low to medium vegetation cover. Conclusions and future scope of the work are discussed in Section VI. II. THEORY Both brightness temperature and radar backscatter have a nearly linear relationship to surface soil moisture, given uniform vegetation and land surface characteristics. The radiative transfer model for estimation of soil moisture from brightness temperature is well established and needs few ancillary parameters for soil moisture estimation. The C-band radiometer AMSR-E has a global soil moisture product and future L-band radiometer such as SMOS and HYDROS will have radiometer-only soil moisture products. However, the radiometer-only soil moisture product is limited in application by the low spatial resolution of the radiometer instrument. Higher spatial resolution is possible with radar soil moisture estimation; however, estimation of absolute soil moisture from radar backscattering coefficients requires modeling a complex signal target interaction. Even in empirical and semiempirical studies, vegetation canopy and soil parameters may be needed to classify a heterogeneous target area into subclasses that are fairly uniform in terms of those parameters. Several studies based on the approach of classification and linear parameterization of L-band radar backscatter measurements with respect to soil moisture within each class have been performed in the past [17], [23]–[25]. The approach taken by the present study is change estimation, which takes advantage of the approximately linear dependence of radar backscatter change on soil moisture change [21]. Njoku et al. demonstrated the feasibility of a change detection approach using the PALS radar and radiometer data obtained during the SGP99 campaign. The PALS and in situ soil moisture data were classified into three different classes based on the vegetation water content. For each class, linear least square fits of PALS brightness temperature and radar backscatter to soil moisture were developed. The linear relationships were modeled as
C
(1) (2)
The PALS data used for the SGP99 study had the same footprint size for both radar and radiometer. Hence, A, B, C, and D are parameters for each pixel in the coincident radar and radiometer images and were assumed to be primarily functions of surface vegetation and roughness (and temperature for the passive case). Difference images were obtained by subtracting the sensor data on the first day from the sensor data on the consecutive days. They were able to calibrate C and D parameters in (2) using two days of radiometer estimates of soil moisture under , they wet and dry soil conditions. Further using C, D, and derived radar estimated soil moisture with satisfactory results. Our study is aimed at estimation of soil moisture change at the spatial resolution of radar by combining radar and radiometer data. The approach and assumptions are similar to the Njoku et al. study; however, in our case the radar is at higher spatial resolution of 100 m as compared to the 400-m spatial resolution of the radiometer. We assume that the changes in vegetation canopy parameters are insignificant as compared to the change in soil moisture when considering the resulting change in copolarized radar backscatter, given a sufficiently high revisit rate of the sensor over the target. Using this assumption, the difference image obtained by subtracting consecutive radar backscatter images acquired over an area would be given by (3) is the change in copolarized radar backscatter (dB) where is the change in soil moisture. The parameter D is exand pected to depend on the attenuation characteristics of the vegetation canopy and the surface roughness characteristics of the soil surface. Results from Du et al. [26] indicate that relative sensitivity of the L-band copolarized channels of radar should depend primarily on the vegetation canopy opacity. Relative sensitivity is defined as the ratio of the radar sensitivity in the presence of a vegetation canopy ( ) to the sensitivity if there was only bare soil (4) For example, Fig. 12 in Du et al. shows the variation of the relative sensitivity for a medium rough soil surface having a soybean canopy. The plot suggests that relative sensitivity can be estimated from optical thickness once the canopy type and surface type are known. The vegetation opacity can be can be esrelationship for vegetation timated using the canopies that follow the electrically thin scatterer approximation [38]. b is a parameter that depends on vegetation structure and type, is the incidence angle, and vwc is the vegetation water content which can be estimated operationally using proxies such as NDWI [34]. Now, combining (3) and (4), we can write (5) depends only weakly on soil roughThe bare soil sensitivity ness variability for a given sensor configuration (frequency, polarization, viewing angle). To substantiate this further, the authors conducted a simulation in which the integral equation model [23] was used to generate plots of vertically copolarized radar backscatter versus soil moisture for various root mean square soil roughness values (Fig. 1). It is seen in the figure 0.4 cm to 2.4 cm can be that the line plots for
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 44, NO. 6, JUNE 2006
1547
scale soil moisture. The summation is for all radar pixels that lie within the radiometer pixel (9) for a particular radiometer pixel can be resampled to the radar spatial resolution and for each radar pixel within the radiometer pixel we write the change in soil moisture at resolution “ ” as (10)
Fig. 1. Simulation of L-band horizontally copolarized radar backscattering coefficients using the integral equation model [23] for various values of and rms surface roughness s (centimeters). volumetric soil moisture Surface correlation length has been taken as 8 cm, sand and clay . For various roughness values, moisture versus backscatter curves can be approximated as straight lines with the same slope. L-band vertically copolarized radar backscattering coefficients show a similar behavior too.
%
= 30
m
%
= 30
approximated as a series of parallel lines with the same slope and different intercepts. The result indicates that for the L-band surface roughness variability has only a small effect on the soil moisture sensitivity of radar. Now, from (5) we obtain (6) where . Let us assume that the radar backscatter is available at a finer resolution of “ ” whereas, radiometer data is available at a coarser resolution of “ .” The problem then is , i.e., the change in soil moisture from time to estimate , given , and at times and step to using (6). The change in soil moisture scales from the higher to lower spatial resolutions by a simple averaging, i.e., (7) The summation is over all the “ ” smaller radar footprints is the change in within the larger radiometer footprint, soil moisture as measured by the radiometer at the lower spatial is the change in soil moisture as measured resolution, by the radar at the higher spatial resolution given by (4). Combining (6) and (7) leads to (8) The unknown will be same for all the radar pixels within a radiometer pixel given uniform vegetation characteristics within the radiometer footprint, i.e., is the same for all radar pixels within the radiometer pixels. The authors recognize the fact that will not be low in a real-world the spatial variability of setting. However, in the case of the SMEX02 experiments, each radiometer footprint lied completely within an agricultural field with fairly uniform vegetation characteristics. In the present study, we do not attempt to model for the vegetation canopy is evaluated for variability within the radiometer footprint. each radiometer pixel by dividing the summation of change in radar backscattering coefficients with the change in radiometer
Another important issue in the low spatial variability of assumption is the implicit assumption that multiple days of radar data were obtained over the region at the same angle of incidence. At different incidence angles, corresponding AIRSAR pixels will exhibit different sensitivity to soil moisture. During SMEX02, the near and far look angles for were 22.8 and 71.3 July 5, 22.0 and 71.2 for July 7 and 24.1 and 71.3 for July 8. This indicates that AIRSAR acquired data over each of the fields at approximately similar incidence angles for the three days. The variation of incidence angles between two fields is not important as the relative change in radar backscatter is considered with the relative change in the soil moisture on a field-wise is derived basis. The low-resolution sensitivity parameter separately for each field. As a result only the variation of incidence angle within each field will be important and this variation is small. Within the dimensions of the PALS footprint, this variation in AIRSAR incident angles will be small and its effect on sensitivity negligible. III. DATA The algorithm discussed above was tested on data obtained from the Soil Moisture Experiments in 2002 (SMEX02). The SMEX02 campaign was conducted in Walnut Creek, a small watershed in Iowa, over a one-month period between mid June and mid July 2002. An extensive dataset of in situ measurements of soil moisture (0–6-cm soil layer), soil temperature (surface, 5-cm depth) soil bulk density, and vegetation water content was collected. Aircraft-mounted instruments—the passive and active, L- and S-band sensor (PALS) and the NASA/JPL airborne synthetic aperture radar (AIRSAR)—were flown with supporting ground-sampling data. The PALS instrument was flown over the SMEX02 region on June 25, 27, and July 1, 2, 5, 6, 7, and 8, 2002 [26], [27]. PALS radiometer and radar provided simultaneous observations of horizontally and vertically polarized L- and S-band brightness temperatures, radar backscatter measured in VV, HH, and VH configurations and thermal infrared surface temperature at a resolution of 400 m. The AIRSAR instrument has P-, L- and C-bands with H/V dual microstrip polarizations and spatial resolutions of 5 m in slant range and 1 m in azimuth [28]. AIRSAR instrument was flown on July 1, 5, 7, 8, and 9. The algorithm proposed in this study needs simultaneous observations of the radar and radiometer. As the PALS spatial coverage of the watershed on July 1 was partial, the data sets for PALS and AIRSAR for July 5, July 7, and July 8 were used. AIRSAR data used in this study was L-band HH and VV polarization and provided at a spatial resolution of 30 m after processing. The AIRSAR images were geolocated
1548
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 44, NO. 6, JUNE 2006
Fig. 2. Plot of change in AIRSAR LVV backscatter at 800-m resolution versus change in in situ volumetric soil moisture at a 800-m resolution for the periods July 5 to July 7 and July 5 to July 8.
Fig. 3. Plot of change in AIRSAR LHH backscatter at 100-m resolution versus change in in situ volumetric soil moisture at 100-m resolution for the periods July 5 to July 7 and July 5 to July 8.
by registration to a Landsat TM7 image. The ground based soil moisture data used in this study was the volumetric soil moisture measured using a theta probe at 14 locations in each field site for all the 31 fields. There were 10 soy fields and 21 corn fields. The m . Seven representative area for each field site was measurements of volumetric soil moisture made along each of two parallel transects, 600 m in length and placed 400 m apart. Higher resolution estimates of in situ soil moisture for each field were estimated by an inverse distance weighted spatial interpolation using all the 14 measurements for each field and a cell size of 100 m. IV. RESULTS As an initial evaluation of radar sensitivity to soil moisture, the AIRSAR L band vertically copolarized backscattering coefficients were aggregated to 800 m and then collocated with the field sites that were sampled for 0–6-cm volumetric soil moisture. Fig. 2 presents a plot of the change in 800-m resolution field averages of AIRSAR LVV backscattering coefficient and soil moisture for the periods July 5 to July 7 and July 5 to July 8. The change in soil moisture between July 7 and July 8 was not very high. In order to see a greater change in soil moisture value of 0.57 the difference July 8–July 5 was selected. The is sensitive to the change in soil moisture; indicates that even under the dense vegetation conditions encountered in the SMEX02 experiments with the vegetation water content of corn fields being around 4–5 kg/m . The sensitivity of the AIRSAR LVV channel to soil moisture was also analyzed at a higher spatial resolution of 100 m. In Figs. 3 and 4, the change in radar backscatter (LHH and LVV, respectively) is compared to the corresponding change in gravimetric soil moisture at a resolution of 100 m for the time pevalues of 0.38 and riods July 5 to July 7 and July 5 to July 8. 0.39 are obtained for LHH and LVV, respectively, indicating that radar sensitivity to soil moisture is significant at the higher spatial resolution of 100 m also. The sensitivities are approximately the same for both LHH and LVV channels with values of 22.2 and 23.8 dB/(cc/cc), respectively. It should be noted that there
Fig. 4. Plot of change in AIRSAR LVV backscatter at 100-m resolution versus change in in situ volumetric soil moisture at 100-m resolution for the periods July 5 to July 7 and July 5 to July 8.
are several data points in Figs. 2–4 with negative backscatter change corresponding to positive change in soil moisture. At the 800-m spatial resolution (Fig. 2), we see data points with a negative change in the range 0 to 1 dB for change in moisture in the range 0 to 0.05 cc/cc. At the 100-m spatial resolution (Figs. 3 and 4), several data points undergo a negative change in the range 0 to 2 dB for change in moisture in the range 0 to 0.1 cc/cc. The authors believe that this effect is primarily due to change in vegetation water content. For example, in Fig. 2, pixels increased in moisture from July 5 and July 7 by a small amount ( 0.05) but still underwent a negative change in backscatter as the vegetation water content increased from July 5 to July 7, causing a greater attenuation of radar backscatter on July 7 as compared to July 5 and resulting in a negative net change in radar backscatter even though the moisture increased. Soil roughness may also change after a rainfall event causing the soil surface to reduce in roughness and result in lower radar backscatter values after the rainfall event. The assumption made by the authors that vegetation and soil roughness do not change between consecutive observations is weak, but it also simplifies
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 44, NO. 6, JUNE 2006
1549
Fig. 5. Difference images for change in PALS LV brightness temperatures at 400-m resolution and change in AIRSAR LVV backscatter at 30-m resolution for the period July 5 to July 7 (July 7–July 5). The spatial patterns corresponding to wetting or drying are strikingly consistent in both images indicating that the AIRSAR LVV channel is sensitive to near surface soil moisture.
the problem significantly with satisfactory results in terms of estimated soil moisture change. It was shown in a previous study that for the SMEX02 field experiment PALS LV channel brightness temperatures were well correlated to soil moisture [29]. A further demonstration of the AIRSAR LVV channel sensitivity to change in soil moisture is done by comparison of the change in PALS LV channel brightness temperatures to the change in AIRSAR LVV channel backscattering coefficients. Fig. 5 presents the difference images produced by the change in AIRSAR LVV backscattering coefficients from July 5 to July 7 compared to the change in PALS LV channel brightness temperature for the same period. The spatial patterns corresponding to wet and dry regions in the watershed are very similar for both the PALS and AIRSAR difference images. Regions that became wetter from July 5 to July 7 underwent a reduction in brightness temperature and an increase in backscattering coefficient. (The scales for the two images in Fig. 5 are, hence, inverted to represent the positive change in radar backscatter and negative change in brightness temperature with an increase in soil moisture.) The correlation between PALS LV channel brightness temperature change and AIRSAR LVV channel radar backscatter change is further brought out in Fig. 6. Change in AIRSAR backscattering coefficients (LVV) have been aggregated to 400-m resolution and compared with the change in PALS brightness temperatures (LV) at its resolution of 400 m. An excellent agreement value of 0.81 indicating is seen between the two with a that AIRSAR LVV channel has a significant sensitivity to soil moisture. The algorithm was applied to three days of AIRSAR LVV, PALS LV, and ground-based soil moisture data. 400-m resoluwere calculated using tion estimates of soil moisture the in situ measurements of soil moisture within each field. As mentioned earlier, 14 theta probe measurements of soil moisture were made at each watershed-sampling site that had dimensions
Fig. 6. Chance in PALS L-band V pol. brightness temperature plotted versus change in AIRSAR LVV channel backscattering coefficeints. AIRSAR data has been aggregated to the PALS resoultion of 400 m. Change is computed for the days July 5 to July 7.
of 800 by 800 m. The in situ measurements were gridded to a 100-m dimension grid using inverse distance interpolation and then they were upscaled to 400 m corresponding to the dimensions of the PALS radiometer footprint. A uniformly distributed random noise of 0–0.016 g/cc was added to the upscaled in situ soil moisture values in order to simulate 4% maximum error that would be obtained if the soil moisture estimates were obtained by inverting soil moisture estimates from L-band brightness temperatures using a radiative transfer model. AIRSAR LVV data was also aggregated to a resolution of 100 m and where “ ” the difference images were used to compute
1550
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 44, NO. 6, JUNE 2006
Fig. 7. The 100-m in situ soil moisture change (x axis) compared with the 100-m resolution estimates of soil moisture change derived from the algorithm for the period July 5 to July 7. Plot 7(a) has all the points with RMSE = 0:046, plot 7(b) has seven outliers removed with RMSE = 0:032.
denotes the lower cell size of 100 m. Using the algorithm discussed in the previous section, 100-m resolution estimates of the change in soil moisture were obtained for two periods of July 5 to July 7 and July 7 to July 8 for each field site. Fig. 7(a) and (b) compares estimated change in soil moisture with measured change in soil moisture at a 100-m resolution for the period July 5 to July 7, and Fig. 8(a) and (b) presents the same comparison for the period July 5 to July 8. The root mean square error for the prediction (RMSE) in both cases is 0.046 cc/cc volumetric, a major portion of which seems to be contributed by a few outliers. It was noted that on removing seven outliers from the plot in Fig. 7(a) and 32 outliers from the plot in Fig. 8(a), the RMSE improves to 0.032 cc/cc and 0.024 cc/cc, respectively. V. SIMULATION EXPERIMENT Only three days of coincident images of AIRSAR and PALS were obtained during the SMEX02 campaign. These data were
Fig. 8. The 100-m in situ soil moisture change (x axis) compared with the 100-m resolution estimates of soil moisture change derived from the algorithm for the period July 5 to July 8. Plot 8(a) has all the points with RMSE = 0:046, plot 8(b) has the data points from fields WC30 and WC31 removed resulting in a RMSE of 0.024.
used to demonstrate the applicability of an approach of spatial disaggregation of soil moisture. This technique was further analyzed in this paper by performing a simulation experiment. Twelve sets of data for soil moisture, vegetation water content, surface roughness, and surface temperature were simulated for a region that is represented by a high-resolution grid consisting of 100 100 cells and a low-resolution grid of 10 10 cells. Each of the datasets were simulated by generating uniformly distributed random numbers within specific ranges (for ) for a 5 5 grid, and example, 0.1–0.4 for soil moisture then spatial interpolation by kriging was done to interpolate the 5 5 grid values to a 100 100 grid. Vegetation was modeled as a soybean canopy with the vegetation water content varying between 0.7 and 0.8 kg/m . Vegetation water content variability was kept low, as the methodology presented in this paper does
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 44, NO. 6, JUNE 2006
1551
Fig. 9. Results from the observation system simulation experiment for the comparison of higher resolution soil moisture disaggregation algorithm with lower resolution estimates obtained from a typical radiometer. Visually it can be seen that the disaggregation algorithm reproduces much of the variability seen in the in situ soil moisture images.
not attempt to model the effect of vegetation water content variis assumed to ability on radar sensitivity to soil moisture ( be uniform within each radiometer footprint). Surface temperature was varied between 20 C and 30 C, and the rms surface roughness value was varied between 1.0 and 2.0 cm. Vegetation water content and surface roughness were held constant for the 12 days of simulated data. A gradual wetting and dry-down was simulated over the 12-day period by allowing the maximum soil moisture for the region to increase from 0.1 on day 0 to 0.45 on day 5 and then decreasing to 0.1 again on day 11. The resulting soil moisture distributions have been presented in Fig. 9 to illustrate the spatial distribution of soil moisture and the simulated wetting and dry-down. Sand and clay percentage for the entire regions was taken as 50% and 30%, respectively, and the bulk density for the entire region was taken as 1.1 g/cc. A forward model for radiative transfer was used to simulate L-band horizontal polarization brightness temperatures [6] at the higher resolution of 100 100 cells and then the simulated brightness temperatures were aggregated to a lower spatial res10 cells. The roughness parameter for the olution of 10 passive case was assumed to vary linearly with and was taken as 0.1 times the surface roughness value at each pixel [38]. Radiometric brightness temperatures were generated at a 10 10 resolution providing for a 10:1 ratio for the radar spatial resolution to the radiometer pixel resolution. For the forward model canopy albedo was taken to be 0 and polarization mixing ratio was assigned a value of 0 [6], [35]. Vegetation opacity was derived from vegetation water content by using a value of
[30]. Zero to 0.3 K 0.138 in relationship noise was introduced in simulations of the brightness temperatures. A listing of parameters used in the radiative transfer model for simulation of L-band brightness temperature has been provided in Table I. A semiempirical backscattering model at the L-band for soybean canopy developed by DeRoo et al., 2001, was used to sim100 resolution ulate backscattering coefficients at the 100 for 12 days [31]. Du et al. [32] used this model in their study to demonstrate that relative sensitivity of radar to soil moisture is primarily a function of vegetation opacity. The present work is also based on this result and, hence, [31] was chosen as the backscattering model. For the bare soil surface model for radar backscatter [31] uses a semiempirical model developed by Oh et al. [16]. The radar backscattering coefficients obtained from the backscattering model were converted to decibels and then uniformly varying random noise in the ranges 0–0.45 dB was added. The low-resolution soil moisture estimate (at 10 10 resolution) was obtained by inverting simulated brightness temperatures using an iterative algorithm to provide a 10 10 image of radiometer predicted soil moistures for all 12 days of the simulation experiment. Using the backscattering model, high-resolution (100 100) datasets for LVV band radar backscatter were also obtained for all 12 days. The algorithm discussed in Section II was applied to the simulated datasets and estimates of soil moisture change were obtained. Algorithm estimate of change at the first time step (Day 2-Day 1) was added to ac-
1552
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 44, NO. 6, JUNE 2006
TABLE I MEDIA AND SENSOR PARAMETERS USED IN THE SIMULATION EXPERIMENT FOR FORWARD MODELING OF BRIGHTNESS TEMPERATURE AND RADAR BACKSCATTER
tual 100 100 resolution soil moisture on the first day leading 100 resolution soil moisture image for to a predicted, 100 the second day. The algorithm estimated change for the second time step (Day 3-Day 2) was added to the predicted 100 100 resolution soil moisture image for Day 2 to give the predicted soil moisture image for Day 3 and so on for all the time steps. The soil moisture predicted (at 100 100 resolution) by the algorithm was compared with actual soil moisture (at 100 100 resolution). The comparison of simulated soil moisture versus the algorithm derived soil moisture for all the days of the simulation experiment has been presented in Fig. 9. The soil moisture maps suggest that the agreement between simulated and algorithm predicted values are very good. It can be seen that the algorithm estimated soil moisture captures much of the spatial variability observed in the simulated soil moisture data sets. The authors would like to point out that the algorithm does not spatially disaggregate absolute soil moisture; rather it spatially disaggregates soil moisture change. Table II presents the root mean squared error between algorithm-estimated change in soil moisture and actual change in soil moisture for each time step. The root mean squared error of prediction for the first time step is 0.004 and the percentage error is 4.8%. The maximum error reached at the sixth time steps is 0.009 volumetric which is around 7.4%. While 7.4% error in relative soil moisture is not acceptable for most applications, it should be noted that the gain in spatial resolution is ten fold. Copolarized radar backscatter–soil moisture relationship was taken to be linear with similar slopes for different surface roughness values. Both the linearity and similarity of slope are first order approximations. However, the authors believe this approximation is only a minor source of error. Major contribution to error comes from a few discrete pixels where the algorithm fails. This
TABLE II ROOT MEAN SQUARED AND PERCENTAGE RETRIEVAL ERRORS ASSOCIATED WITH THE ALGORITHM RETRIEVED SOIL MOISTURE FOR 12 DAYS OF SIMULATION
Fig. 10. Frequency distribution for the change estimation error at each pixel for the time step from Day 3 to Day 4. Pixels having error around the 0.06 cc/cc value are shown in the estimated soil moisture image for Day 4.
is due to two factors, the change in soil moisture as predicted by the radiometer is close to zero or the averaged change in radar backscatter becomes close to zero causing the sensitivity computation to be numerically unstable. This effect is seen in Fig. 10, which shows the frequency distribution for change estimation errors for the time step Day 3 to Day 4. A bimodal frequency distribution is seen with modes around 0 and 0.06. The inset shows the pixels that are contributing to this error. As just mentioned this error seems to be a result of numerical instability during sensitivity computation. One way of avoiding this error in an operational case would be doing the sensitivity computation not for each time step, but choosing the most wet and most dry condition for a time period of 7–8 days. Vegetation is not expected to change significantly in that period and, hence, wettest and driest conditions observed during that period may allow for a more stable sensitivity computation. While the higher resolution algorithm estimates of soil moisture produce much of the soil moisture variability seen in the simulated soil moisture images, the estimated soil moisture from
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 44, NO. 6, JUNE 2006
simulated brightness temperature does not produce a similar spatial variability due to the inherent problem of lower spatial resolution. A good feature of this algorithm presented in this paper is that if the higher resolution spatially disaggregated images of soil moisture are aggregated to the resolution of the radiometer the aggregated soil moisture images will be same as those produced by the radiometer and as a result the radiometer and radar/radiometer (algorithm) estimates of soil moisture will be consistent. VI. CONCLUSION This paper proposed and implemented a simple algorithm for estimation of change in soil moisture at the spatial resolution of radar using low-resolution estimate of soil moisture from radiometer and copolarized backscattering coefficients. The authors argue that subpixel scale surface roughness variability does not play an important role in radar sensitivity to soil moisture at the L-band. Observations from combined radar/radiometer data from SMEX02, results from previous studies and IEM simulations have been presented in support of this argument. Radar sensitivity to soil moisture at the L-band has been assumed to be a function of vegetation opacity only and further a simple soil moisture change estimation algorithm has been developed. Application of the algorithm to data obtained from the SMEX02 experiments gave excellent results with root mean square error of prediction of 0.03 and 0.02 (error for estimated versus measured volumetric soil moisture, both at 100-m resolution) for two periods—July 5 to July 7 and July 7 value for both cases was 0.85. A simulation to July 8. The experiment was performed to examine the performance of the algorithm over a range of ground parameter values (soil moisture, surface roughness, surface temperature). The range of root mean square errors associated with estimation of soil moisture change was 0.004 to 0.009 volumetric soil moisture which or 4.8% to 7.4%. It should be noted that the simulation algorithm assumes that parameters such has surface roughness, vegetation opacity, and surface temperature are known accurately and this will not be true in an operational setting. It is assumed in the present study that the 0.45-dB and 0.3-K uncertainties added simulated radar backscatter and brightness temperatures in the form of random noise would take into account some of the parametric uncertainty. However, a more rigorous evaluation of the approach presented in the paper with uncertainty incorporated into all physical and remote sensing measurements and the impact of these uncertainties is beyond the scope of this study. The originality of the approach presented here lies in using radiometer to estimate soil moisture change at a lower spatial resolution (but with lower ancillary data requirements as compared to radar estimation of soil moisture) and then using change in radar backscatter to estimate the change in soil moisture at higher spatial resolution. The estimated change in soil moisture is a hydrologic variable of significant interest. A simple calibration methodology based on least error between modeled and measured soil moisture change values will allow a better estimation of parameters such as the hydraulic conductivity of soil layers. Mattikalli et al. reported that the two-days soil moisture
1553
change was closely related to the saturated hydraulic conducprofile [33]. It will be possible to relate K from tivity K the radar/radiometer algorithm derived change in soil moisture with the added advantage of higher spatial resolution that will lead to more accurate estimation of water and energy fluxes. Future studies the direction of radar/radiometer combination will have to aim at a better parameterization of the sensitivity relationship with vegetation opacity allowing the effect of vegeparamtation heterogeneity to be addressed through the eter in the algorithm. Du et al. have explored the behavior of soybean and grass canopies over medium rough surfaces [32]. Their results indicate that it should be possible to develop simple parametric relationships between vegetation opacity and relative sensitivity. Estimation of vegetation opacity has been done in the past using optical sensors by estimation of vegetation water content using a proxy such as NDWI [34] and then using the empirical parameter to relate vegetation opacity and water content [30]. This simple parameterization, however, has been shown to be too simplistic for canopies such as corn that are electrically thick scatterers and further research is required to find relationships between vegetation parameters and relative sensitivity over canopies such as corn. The approach presented in the paper should be applicable to data from the HYDROS mission, at least over areas of low-vegetation water content variability. The algorithm will have to be modified to account for vegetation variability within the radiometer footprint using the parameter before it can be made operational. ACKNOWLEDGMENT The authors would like to thank J. Van Zyl (Jet Propulsion Laboratory) for providing the processed AIRSAR data used in this study. They would also like to thank M. Cosh for providing the Landsat landcover and vegetation water content data. Finally, the authors would like to thank E. N. Njoku for providing guidance during the course of this study. REFERENCES [1] J. Shukla and Y. Mintz, “Influence of land-surface evapo-transpiration on the earths climate,” Science, vol. 215, no. 4539, pp. 1498–1501, 1982. [2] T. Delworth and S. Manabe, “The influence of potential evaporation on the variablities of the simulated soil wetness and climate,” J. Climate, vol. 1, pp. 523–547, 1988. [3] T. J. Jackson, T. J. Schmugge, and E. T. Engman, “Remote sensing applications to hydrology: soil moisture,” Hydrol. Sci. J.-J. Des Sci. Hydrol., vol. 41, no. 4, pp. 517–530, 1996. [4] M. J. Fennessy and J. Shukla, “Impact of initial soil wetness on seasonal atmospheric prediction,” J. Clim., vol. 12, no. 11, pp. 3167–3180, 1999. [5] P. A. Dirmeyer, F. J. Zeng, A. Ducharne, J. C. Morrill, and R. D. Koster, “The sensitivity of surface fluxes to soil water content in three land surface schemes,” J. Hydrometeorol., vol. 1, no. 2, pp. 121–134, 2000. [6] E. G. Njoku and D. Entekhabi, “Passive microwave remote sensing of soil moisture,” J. Hydrol., vol. 184, no. 1–2, pp. 101–129, 1996. [7] F. T. Ulaby, P. C. Dubois, and J. vanZyl, “Radar mapping of surface soil moisture,” J. Hydrol., vol. 184, no. 1–2, pp. 57–84, 1996. [8] T. J. Schmugge, W. P. Kustas, J. C. Ritchie, T. J. Jackson, and A. Rango, “Remote sensing in hydrology,” Adv. Water Resources, vol. 25, no. 8–12, pp. 1367–1385, 2002. [9] M. Owe, R. de Jeu, and J. Walker, “A methodology for surface soil moisture and vegetation optical depth retrieval using the microwave polarization difference index,” IEEE Trans. Geosci. Remote Sens., vol. 39, no. 8, pp. 1643–1654, Aug. 2001. [10] E. G. Njoku, T. J. Jackson, V. Lakshmi, T. K. Chen, and S. V. Nghiem, “Soil moisture retrieval from AMSR-E,” IEEE Trans. Geosci. Remote Sens., vol. 41, no. 2, pp. 215–229, Feb. 2003.
1554
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 44, NO. 6, JUNE 2006
[11] J. R. Wang, J. C. Shiue, T. J. Schmugge, and E. T. Engman, “The L-band Pbmr measurements of surface soil-moisture in fife,” IEEE Trans. Geosci. Remote Sens., vol. 28, no. 5, pp. 906–914, May 1990. [12] T. J. Schmugge, T. J. Jackson, W. P. Kustas, and J. R. Wang, “Passive microwave remote-sensing of soil-moisture—Results from hapex, fife and Monsoon-90,” ISPRS J. Photogramm. Remote Sens., vol. 47, no. 2–3, pp. 127–143, 1992. [13] P. E. O’Neill and N. Chauhan, “Use of active and passive microwave remote sensing for soil moisture estimation through corn,” Int. J. Remote Sens., vol. 17, pp. 1851–1865, 1996. [14] T. J. Jackson, D. M. Levine, C. T. Swift, T. J. Schmugge, and F. R. Schiebe, “Large-area mapping of soil-moisture using the estar passive microwave radiometer in Washita92,” Remote Sens. Environ., vol. 54, no. 1, pp. 27–37, 1995. [15] A. P. Barros, R. Bindlish, and A. Rogowski, “Soil Hydrology and Spatial Variability-Perspectives on the Interpretation of Remotely Sensed Data,” Pennsylvania State Univ., Environ. Inst., University Park, vol. 65, 2000. Tech. Rep. Series. [16] Y. K. Oh, K. Sarabandi, and F. T. Ulaby, “An empirical model and an inversion technique for radar scattering from bare surfaces,” IEEE Trans. Geosci. Remote Sens., vol. 30, no. 2, pp. 370–381, Mar. 1992. [17] P. C. Dubois, J. J. van Zyl, and E. T. Engman, “Measuring soil moisture with imaging radars,” IEEE Trans. Geosci. Remote Sens., vol. 33, no. 4, pp. 915–926, Jul. 1995. [18] J. C. Shi, J. Wang, A. Y. Hsu, P. E. O’Neill, and E. T. Engman, “Estimation of bare surface soil moisture and surface roughness parameter using L-band SAR image data,” IEEE Trans. Geosci. Remote Sens., vol. 35, no. 5, pp. 1254–1266, Sep. 1997. [19] M. Owe, A. A. Van de Griend, R. de Jeu, J. J. de Vries, E. Seyhan, and E. T. Engman, “Estimating soil moisture from satellite microwave observations: past and ongoing projects, and relevance to GCIP,” J. Geophys. Res.-Atmos., vol. 104, no. D16, pp. 19735–19 742, 1999. [20] E. T. Engman and N. Chauhan, “Status of microwave soil-moisture measurements with remote-sensing,” Remote Sens. Environ., vol. 51, no. 1, pp. 189–198, 1995. [21] E. G. Njoku, W. J. Wilson, S. H. Yueh, S. J. Dinardo, F. K. Li, T. J. Jackson, V. Lakshmi, and J. Bolten, “Observations of soil moisture using a passive and active low-frequency microwave airborne sensor during SGP99,” IEEE Trans. Geosci. Remote Sens., vol. 40, no. 12, pp. 2659–2673, Dec. 2002. [22] J. D. Villasenor, D. R. Fatland, and L. D. Hinzman, “Change detection on Alaska’s North Slope using repeat-pass ERS-1 SAR images,” IEEE Trans. Geosci. Remote Sens., vol. 31, no. 1, pp. 227–236, Jan. 1993. [23] K. Fung, Z. Li, and K. S. Chen, “Backscattering from a randomly rough dielectric surface,” IEEE Trans. Geosci. Remote Sens., vol. 30, no. 2, pp. 356–359, Mar. 1992. [24] M. C. Dobson and F. T. Ulaby, “Active microwave soil-moisture research,” IEEE Trans. Geosci. Remote Sens., vol. GRS-24, no. 1, pp. 23–36, Jan. 1986. [25] M. C. Dobson and F. T. Ulaby, “Preliminary evaluation of the Sir-B response to soil-moisture, surface-roughness, and crop canopy cover,” IEEE Trans. Geosci. Remote Sens., vol. GRS-24, no. 4, pp. 517–526, Apr. 1986. [26] M. Cosh, T. J. Jackson, R. Bindlish, and J. Prueger, “Watershed scale temporal persistence of soil moisture and its role in validating satellite estimates,” Remote Sens. Environ., vol. 92, pp. 427–435, 2004. [27] A. S. Limaye, W. L. Crosson, C. A. Laymon, and E. G. Njoku, “Land cover-based optimal deconvolution of PALS L-band microwave brightness temperatures,” Remote Sens. Environ., vol. 92, pp. 497–506, 2004. [28] L. Yunling, K. Yunjin, and J. van Zyl. The NASA/JPL Airborne Synthetic Aperture Radar System. [Online] Available: http://airsar.jpl.nasa.gov/documents/genairsar/airsar_paper1.pdf [29] U. Narayan, V. Lakshmi, and E. G. Njoku, “Retrieval of soil moisture from passive and active L/S band sensor (PALS) observations during the Soil Moisture Experiment in 2002 (SMEX02),” Remote Sens. Environ., vol. 92, no. 4, pp. 483–496, 2004. [30] T. J. Jackson and T. J. Schmugge, “Vegetation effects on the microwave emission of soils,” Remote Sens. Environ., vol. 36, pp. 203–212, 1991. [31] R. D. De Roo, Y. Du, F. T. Ulaby, and M. C. Dobson, “A semi-empirical backscattering model at L-band and C-band for a soybean canopy with soil moisture inversion,” IEEE Trans. Geosci. Remote Sens., vol. 39, no. 4, pp. 864–872, Apr. 2001. [32] Y. Du, F. T. Ulaby, and M. C. Dobson, “Sensitivity to soil moisture by active and passive microwave sensors,” IEEE Trans. Geosci. Remote Sens., vol. 38, no. 1, pp. 105–114, Jan. 2000.
[33] N. M. Mattikalli, E. T. Engman, T. J. Jackson, and L. R. Ahuja, “Microwave remote sensing of temporal variations of brightness temperature and near-surface soil water content during a watershed-scale field experiment, and its application to the estimation of soil physical properties,” Water Resources Res., vol. 34, no. 9, pp. 2289–2299, 1998. [34] T. J. Jackson, D. Chen, M. Cosh, F. Li, M. Anderson, C. Walthall, P. Doraiswamy, and E. Hunt, “Vegetation water content mapping using Landsat data derived normalized difference water index for corn and soybeans,” Remote Sens. Environ., vol. 92, no. 4, pp. 475–482, 2004.
Ujjwal Narayan received the B.S. degree in civil engineering from the Indian Institute of Technology, Kanpur, India, in 1998 and is currently pursuing the Ph.D. degree in geological sciences at the University of South Carolina, Columbia. His research interests include interaction of microwave emission and backscattering from soil surfaces under vegetation canopies, synergistic use of microwave and optical data for estimation of soil moisture, and assimilation of soil moisture information at different spatial and temporal scales into a hydrologic model. Mr. Narayan is a student member of the American Geophysical Union and the IEEE Geoscience and Remote Sensing Society. He received the NASA Graduate Student Fellowship for Earth Science in 2005.
Venkataraman Lakshmi (S’94–M’99–SM’02) received the B.S. degree in civil engineering from the University of Roorkee, Roorkee, India, in 1987, the M.S. degree in civil and environmental engineering from the University of Iowa, Iowa City, in 1989, and the M.A. and Ph.D. degrees in civil engineering and operations research, respectively, from Princeton University, Princeton, NJ, in 1995. He has been working in the remote sensing and modeling of the land surface hydrological cycle for the past 12 years on various aspects. These range from catchment hydrology to continental-scale hydrology and climate using data from aircraft sensors as well as satellites. He is currently an Assistant Professor in the Department of Geological Sciences, University of South Carolina, Columbia. Dr. Lakshmi is the Chairman of the AGU remote sensing committee for the hydrology section, Associate Editor for J. Geophysical Research, and editor for hydrology for the American Geophysical Union publication EOS. He is a member of the IEEE Geoscience and Remote Sensing Society Technical Committees on Data Fusion and New Technologies as well as serving on the Technical Paper committee. He is a member of the AMSR Validation team (NASA and NASDA) as well as the EOS IDS Working Group.
Thomas J. Jackson (A’86–SM’96–F’02) received the Ph.D. degree in civil engineering from the University of Maryland, College Park, in 1976. He is currently a Hydrologist with the USDA Agricultural Research Service Hydrology and Remote Sensing Laboratory, Beltsville, MD. His research involves the application and development of remote sensing technology in hydrology and agriculture. He has conducted research on the use of visible/near-infrared satellite data for deriving land-cover parameters used in hydrologic models and the use of an airborne laser profiler for measuring and monitoring soil erosion. His current research focuses on the use of passive microwave techniques in hydrology. These studies have ranged from small-scale controlled condition field experiments utilizing truck-mounted radiometers to large scale multitemporal aircraft mapping. Dr. Jackson received The Paper of the Year Award from the American Society of Agricultural Engineers for his paper “Airborne Laser Measurements of the Surface Topography of Simulated Concentrated Flow Gullies” in 1990 and the prize paper award for his “Diurnal Observations of Surface Soil Moisture Using Passive Microwave Radiometers” at the International Geoscience and Remote Sensing Symposium in 1994. He is a Fellow of the American Geophysical Union.
