Next Article in Journal
The Emergence and Spread of Ecourban Neighbourhoods around the World
Next Article in Special Issue
Multi-Scale Measurement of Regional Inequality in Mainland China during 2005–2010 Using DMSP/OLS Night Light Imagery and Population Density Grid Data
Previous Article in Journal
Sustainable Modernization of the Russian Power Utilities Industry
 
 
Due to planned maintenance work on our platforms, there might be short service disruptions on Saturday, December 3rd, between 15:00 and 16:00 (CET).
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Temperature Vegetation Dryness Index Estimation of Soil Moisture under Different Tree Species

1
College of Economics and Management, Nanjing Forestry University, Nanjing 210000, China
2
International Institute for Earth System Science, Nanjing University, Nanjing 210000, China
*
Author to whom correspondence should be addressed.
Sustainability 2015, 7(9), 11401-11417; https://doi.org/10.3390/su70911401
Received: 1 July 2015 / Revised: 11 August 2015 / Accepted: 12 August 2015 / Published: 25 August 2015

Abstract

:
The Laoshan forest is the largest forest in Nanjing, and it plays an important role in water resource management in Nanjing. The objectives of this study are to determine if the temperature vegetation dryness index (TVDI) is suitable to estimate the soil moisture and if soil moisture is significantly affected by tree species in the Laoshan forest. This paper calculated the spatial distribution of TVDI using LANDSAT-5 TM data. Sixty-two observation points of in situ soil moisture measurements were selected to validate the effectiveness of the TVDI as an index for assessing soil moisture in the Laoshan forest. With the aid of the three different temporal patterns, which are 10 January 2011, 18 May 2011 and 23 September 2011, this paper used the TVDI to investigate the differences of soil moisture under four kinds of mono-species forests and two kinds of mixed forests. The results showed that there is a strong and significant negative correlation between the TVDI and the in situ measured soil moisture (R2 = 0.15–0.8, SE = 0.015–0.041 cm3/cm3). This means that the TVDI can reflect the soil moisture status under different tree species in the Laoshan forest. The soil moisture under these six types of land cover from low to high is listed in the following order: Eucommia ulmoides, Quercus acutissima, broadleaf mixed forest, Cunninghamia lanceolata, coniferous and broadleaf mixed forest and Pinus massoniana.

1. Introduction

Soil moisture is a very important variable of the climate system, as it controls numerous processes and feedback loops within the climate system. It is of major relevance for the global water, energy and carbon cycles [1]. Increased climate change is altering the global water cycle and affecting the amount of water available for tree species. Differences in soil water content of tree species are becoming increasingly important to identify with global climate changes [2]. The Laoshan forest is the largest forest in Nanjing. In addition, with the rapid growth of the economy and urban population, the problem of supply and demand for water in Nanjing has become increasingly important over the past several decades [3]. Therefore, knowledge of the state of soil moisture for different tree species over time in the Laoshan forest is essential for a wide range of meteorological and hydrological applications, such as weather and climate prediction, terrestrial carbon cycle simulation, water management and policy planning.
There are many different techniques for in situ soil moisture measurement under tree species, such as the gravimetric method, neutron probe, heat dissipation sensor, time domain reflectometry (TDR), frequency domain reflectometry (FDR), etc. [4,5]. However, in situ soil moisture measurements are cost intensive and require major efforts to be put in place. As a result, only a few in situ soil moisture measurement networks are available. In addition, it is often questioned whether the point measurements for regional applications are representative [6,7].
Remote sensing is currently in a strong position to provide meaningful spatial and temporal data for use in soil moisture investigations. Over the past 40 years, substantial research has been carried out to retrieve soil moisture using remotely-sensed observations [8], but little research focused on investigating the differences of soil moisture under different tree species. Therefore, this paper investigates the suitability of remote sensing for estimating the soil moisture of different tree species in the Laoshan forest. In this study, the accuracy of estimated soil water content under different tree species is investigated with the aid of three different temporal patterns.
The normalized difference vegetation index (NDVI) can monitor vegetation status and stress, specifically in relation to water stress, and the forest canopy temperature (Ts) will rise rapidly with water stress [9,10]. The potential for obtaining soil moisture through the relationship between remotely-sensed Ts and NDVI has been investigated by several authors [9,10,11]. The slope of the Ts/NDVI curve is related to the evapotranspiration rate of the surface and is used to assess soil moisture conditions. Numerous studies focus on the slope of the Ts/NDVI curve for this purpose [12]. The scatter plots of Ts/NDVI space often result in a triangular shape [10,13] or a trapezoidal shape [14,15].
Sandholt et al. [10] explored a simplified land surface dryness index, called the temperature vegetation dryness index (TVDI), based on an empirical parameterization of the relationship between Ts and NDVI. Following this, different satellite images have been used to demonstrate the potential of the TVDI for soil moisture estimation, such as the National Oceanic and Atmospheric Administration (NOAA) Advanced Very High Resolution Radiometer (AVHRR) images [10,16], Terra/Aqua Moderate-Resolution Imaging Spectroradiometer (MODIS) images [7,8,17,18,19,20] and LANDSAT-5 Thematic Mapper (TM) images [21]. The comparisons with soil moisture in situ measurements or model simulations show that the TVDI is feasible for monitoring soil moisture.
The scale effect is one of the very important scientific problems of remote sensing [22], and it can increase uncertainty in soil moisture retrieval. LANDSAT-5 TM images have a higher spatial resolution (30–120 m) than Terra/Aqua MODIS (250–1000 m) images or NOAA-AVHRR (1100 m) images, so this study uses LANDSAT-5 TM images.
There are two objectives of this study: (i) to determine the suitability of TVDI for estimating the soil moisture in the Laoshan forest; and (ii) to determine if soil moisture is significantly affected by tree species. Three different temporal patterns of satellite images, 10 January 2011, 18 May 2011 and 23 September 2011, were selected in this study.

2. Study Area

The Laoshan forest is located in Nanjing Jiangsu Province in East China between 32°02′34ʺ and 32°09′54ʺ N latitude and 118°24′33ʺ and 118°41′15ʺ E longitude (Figure 1). Its range is 35 km from east to west, and 15 km from south to north. The total area is 7493 ha. There are some water reservoirs in the Laoshan forest, and thematic maps of them were generated from the forest inventory organized by the Jiangsu Provincial Forestry Bureau in 2007.
Figure 1. The location of the Laoshan forest in Nanjing. (a) The location of the Laoshan forest in Nanjing and (b) the types of land cover in the Laoshan forest region. The yellow points in (b) are the sixty-two observation points of in situ soil moisture measurements.
Figure 1. The location of the Laoshan forest in Nanjing. (a) The location of the Laoshan forest in Nanjing and (b) the types of land cover in the Laoshan forest region. The yellow points in (b) are the sixty-two observation points of in situ soil moisture measurements.
Sustainability 07 11401 g001
There are four kinds of mono-species and two kinds of mixed forests in the Laoshan forest. The tree species of these mono-species forests are Cunninghamia lanceolata, Eucommia ulmoides, Quercus acutissima and Pinus massoniana. The two kinds of mixed forests are broadleaf mixed forest and coniferous and broadleaf mixed forest. There are three different types of tree species in the broadleaf mixed forest, and they are Quercus acutissima, Celtis sinensis and Sophora japonica. The coniferous and broadleaf mixed forest mainly consists of the following five different types of tree species, which are Quercus acutissima, Pterocarya stenoptera, Celtis sinensis, Pinus massoniana and exotic pine. The areas of each land cover type in the Laoshan forest are presented in Table 1.
Table 1. The areas of each land cover type in the Laoshan forest.
Table 1. The areas of each land cover type in the Laoshan forest.
Land Cover TypeArea (units: ha)
broadleaf mixed forest5364.988
Quercus acutissima704.342
Pinus massoniana606.933
coniferous and broadleaf mixed forest359.664
Eucommia ulmoides269.748
Cunninghamia lanceolata187.325

3. Methodology

This study uses a land surface dryness index called TVDI to estimate the soil moisture under different tree species in the Laoshan Forest. Figure 2 presents a flowchart of the soil moisture estimation of different tree species assessment. All steps will be discussed in detail below.
Figure 2. The flowchart of the soil moisture under different tree species’ assessment.
Figure 2. The flowchart of the soil moisture under different tree species’ assessment.
Sustainability 07 11401 g002

3.1. Field Work

In order to carry out the analysis for different vegetation covers and different temporal patterns, sixty-two observation points of in situ soil moisture measurements were selected to ensure that each tree species had at least five observation points (see Figure 1). The method of TVDI is based on the feedbacks of soil moisture content to the forest canopy temperature. The soil moisture content within the root zone affects the plants’ transpiration. With decreasing soil moisture content, the soil suction increases, the remaining soil moisture becomes less available for uptake by plant roots, the transpiration might thus become reduced and, then, the forest canopy temperature will increase [1]. Previous results showed that in the root zone of the soil profile, the relationships between the TVDI and soil moisture at depths of 0–20 cm of soil samples were closer than other depths [7,17,19,23,24]. Based on these previous results, for this study, volumetric soil samples (the volume of each soil sample is 100 cm3) were collected 0–15 cm deep. The sample mass and volume were determined before and after oven drying (i.e., generally drying sample in an oven set at 105 degrees Celsius for 12 h) to determine gravimetric water content, bulk density and, hence, volumetric water content.

3.2. Satellite Image Processing

In this study, image rectification, including geometric rectification, radiometric calibration, solar zenith angle correction and atmospheric correction, has been accomplished using ENVITM image processing software (Version 4.7). Radiometric calibration of Landsat-5 TM is based on the analysis of Chander et al. [25]. Atmospheric correction of Landsat-5 TM is accomplished using the ENVI FLAASH (Fast Line-of-sight Atmospheric Analysis of Hypercubes, FLAASH) model which is based on the MODTRAN4 RT model [26].

3.3. Calculation of Normalized Difference Vegetation Index

NDVI is the observed normalized difference vegetation index, and it can be defined as:
NDVI = (ρnirρred)/(ρnir + ρred)
where ρnir is the near-infrared band reflectance and ρred is the red band reflectance. In this study, NDVI was calculated using LANDSAT-5 TM Band 3 (the red band) reflectance and LANDSAT-5 TM Band 4 (the near-infrared band) reflectance.

3.4. Retrieval of Surface Temperature (Ts)

Qin et al. [27] developed a mono-window algorithm for retrieving surface temperature from LANDSAT-5 TM Band 6 data. Many researchers have validated the algorithm and have shown that the algorithm provides an RMSD value of 0.9–1.1 K [28,29,30]. Thus, this study uses this algorithm to obtain the Ts. The MODIS land surface temperature products (MOD11A1, Version 5) were used to validate the Ts. These data have been downloaded from the Land Processes Distributed Active Archive Center (https://lpdaac.usgs.gov/). The spatial resolution of the Ts products (MOD11A1) is 1 km; however, the spatial resolution of Ts that we retrieved is 120 m. In order to correspond to the spatial resolution of the MOD11A1 Ts products, this paper calculated the mean Ts of the 8 × 8 pixels. The validation results are shown in Figure 3. The R2 and RMSE are 0.74–0.85 and 2.5–2.9 K, respectively. This means that the results of estimated Ts are acceptable.
Figure 3. Comparison of the estimated Ts with the MODIS Ts products on: (a) 10 January 2011; (b) 18 May 2011; and (c) 23 September 2011.
Figure 3. Comparison of the estimated Ts with the MODIS Ts products on: (a) 10 January 2011; (b) 18 May 2011; and (c) 23 September 2011.
Sustainability 07 11401 g003

3.5. Calculation of Temperature Vegetation Dryness Index

TVDI is a simplified land surface dryness index, which is based on an empirical formula of the relationship between Ts and NDVI, and only satellite-derived information is used in the method [10]. Therefore, the TVDI was used to estimate the soil moisture in the Laoshan forest. The scatter plots of the Ts/NDVI space often result in a trapezoidal shape [14,15]. Figure 4 shows the conceptual Ts/NDVI space.
Figure 4. Definition of the TVDI. The TVDI for a given NDVI is estimated using Ts, Tsmin and Tsmax (see Equation (2)).
Figure 4. Definition of the TVDI. The TVDI for a given NDVI is estimated using Ts, Tsmin and Tsmax (see Equation (2)).
Sustainability 07 11401 g004
In the trapezium, the upper sloping edge of the trapezium is defined as the dry edge, and the lower sloping edge is defined as the wet edge; they represent extreme conditions of soil moisture and evapotranspiration. The points closer to the dry edge reflect a much stressed surfaces, with lower soil moisture in the root zone and higher Ts in the surface [17]. On the dry edge, as the NDVI increases along the x-axis, the maximum Ts decreases, and stomatal resistance to evapotranspiration is a key factor, which is partly controlled by the limited moisture availability [10]. On the other hand, the wet edge consists of a group of points forming a horizontal line. On the wet edge, under no-water-stress conditions, the Ts is independent of the NDVI. For the points closer to the wet edge, the evapotranspiration capacity and soil moisture become higher.
For a given pixel, Tsmax are the values of Ts on the dry edge, respectively, for the value of the NDVI for that pixel. The TVDI is defined as [10]:
TVDI = (TsTsmin)/( TsmaxTsmin)
and it can be determined on a pixel-by-pixel basis. On the dry edge, Tsmax can be represented by straight line relations with the NDVI, i.e., as:
Tsmax = a + b(NDVI)
where the coefficients a and b can be determined by a least squares fit to the actual data.
In this paper, we used LANDSAT-5 TM images (Path 120, Row 32) for three different times of the year, 10 January 2011, 18 May 2011 and 23 September 2011, to determine the NDVI, Ts, Tsmin and Tsmax values.

4. Results and Discussion

4.1. Analysis of the Ts/NDVI Feature Space

The plots of Tsmin and Tsmax as a function of NDVI for the three scenes are shown in Figure 5. The peak values of the dry edge are in the bare soil. With the increase of NDVI values, the peak values of the dry edge reduce because of the high rate of evapotranspiration from vegetation canopies. The amount of annual average rainfall in Nanjing is 1090.6 mm; approximately 55% of the annual rainfall is between May and August. Therefore, the soil water is sufficient for evapotranspiration by vegetation canopies on 18 May 2011 and 23 September 2011, and Tsmin decreases in line with NDVI increasing. There was precipitation on 17 September 2011, and the amount of rainfall was 16 mm; therefore, the slope of the dry edge and the wet edge on 23 September 2011 is greater than that on 18 May 2011. On 10 January 2011, Tsmin increases slowly in line with NDVI increasing. The reason for this phenomenon is that the precipitation of Nanjing from the autumn of 2010 to the spring of 2011 was less than normal and caused continuous drought [31]. As a result, the soil water supply cannot fully meet the needs for vegetation evapotranspiration in the forest with vegetation cover increasing, and the slope of the dry edge on 10 January 2011 is less than that on 18 May 2011 and 23 September 2011.
Figure 5. Plots of NDVI (x-axis) and surface temperature-Ts (y-axis) in the study. The maximum Ts and minimum Ts are extracted for small intervals (0.01) of NDVI.
Figure 5. Plots of NDVI (x-axis) and surface temperature-Ts (y-axis) in the study. The maximum Ts and minimum Ts are extracted for small intervals (0.01) of NDVI.
Sustainability 07 11401 g005aSustainability 07 11401 g005b
Likewise, Tsmax is negatively correlated with NDVI and can be obtained through linear regression. In order to determine the parameters describing the dry and wet edges, the maximum and minimum temperatures observed for small intervals of NDVI are extracted in the Ts/NDVI space. The values of the coefficients in Equation (3) for the dry edges and wet edges for the three scenes have been determined and are given in Table 2.
Table 2. Dry edges and wet edges in the Ts/NDVI space for the three scenes.
Table 2. Dry edges and wet edges in the Ts/NDVI space for the three scenes.
DateDry EdgeWet Edge
10 January 2011Tsmax = –4.049 (NDVI) + 13.037
R² = 0.72
Tsmin = 0.819 (NDVI) + 1.726
R² = 0.26
18 May 2011Tsmax = –5.510 (NDVI) + 29.621
R² = 0.82
Tsmin = –2.514 (NDVI) + 12.387
R² = 0.46
23 September 2011Tsmax = –10.285 (NDVI) + 29.029
R² = 0.93
Tsmin = –4.734(NDVI) + 13.942
R² = 0.59

4.2. Spatial Variation of TVDI

The TVDI of the Laoshan forest for the three different scenes has been calculated, and it is shown in Figure 6.
On 23 September 2011, there was a small amount of cloud in the sky. The region covered by cloud was screened out and is represented in black in Figure 6c. The area with low TVDI values on 23 September 2011 was larger than that on 18 May 2011. The mean of TVDI on 23 September 2011 was 0.56, and the mean of TVDI on 18 May 2011 was 0.60. However, the area with low TVDI values on 18 May 2011 was larger than that on 10 January 2011, and the mean of TVDI on 10 January 2011 was 0.74. The reason for this phenomenon is that the soil moisture is high between May and August in Nanjing, so the soil water is sufficient for evapotranspiration by vegetation canopies on 18 May 2011 and 23 September 2011. There was precipitation on 17 September 2011, so the TVDI on 23 September 2011 was lower than that on 18 May 2011. Because of the continuous drought from the autumn of 2010 to the spring of 2011, the TVDI on 10 January 2011 was higher than that on 23 September 2011 and 18 May 2011. There are some reservoirs in the Laoshan forest (see Figure 1), so the areas near the reservoirs have very low TVDI values, and the values are less than 0.5.
Figure 6. Spatial distribution of the TVDI for the three scenes. On 23 September 2011, the area covered by cloud was screened out and is represented in black in Figure 6c.
Figure 6. Spatial distribution of the TVDI for the three scenes. On 23 September 2011, the area covered by cloud was screened out and is represented in black in Figure 6c.
Sustainability 07 11401 g006

4.3. Comparison of TVDI with In Situ Measurements

The effectiveness of the TVDI as an index for assessing soil moisture was validated by systematically designed in situ soil moisture measurements for each land cover type (see Figure 1). Figure 7 shows the relationship between TVDI and the volumetric soil moisture at the observation points.
Figure 7. Comparison of the TVDI with the observation points of in situ soil moisture measurements for each land cover on 10 January 2011, 18 May 2011 and 23 September 2011.
Figure 7. Comparison of the TVDI with the observation points of in situ soil moisture measurements for each land cover on 10 January 2011, 18 May 2011 and 23 September 2011.
Sustainability 07 11401 g007aSustainability 07 11401 g007b
On 23 September 2011, eleven observation points were of no use for validation purposes because of cloud, so only fifty-one observation points exist in Figure 7. Linear relationships exist between soil moisture and TVDI, and the coefficients in the equation of the regression line:
Mv = e(TVDI) + f
where Mv is the volumetric soil moisture, in cm3/cm3, the coefficients e and f can be determined by a least squares fit to the actual data and are given in Table 3 for the three different scenes that we have analyzed.
In Figure 7, soil moisture values plotted as a function of TVDI show higher TVDI values corresponding to lower soil moisture values. There is a significant negative correlation between the TVDI and the in situ measured soil moisture, and the standard errors of soil moisture estimation (SE) are low (SE = 0.015–0.041 cm3/cm3). Therefore, this means that the TVDI can reflect the soil moisture status. The relationships between soil moisture and TVDI are closer on 18 May 2011 and 23 September 2011 (R2 = 0.29–0.8 and SE = 0.015–0.041 cm3/cm3) than on 10 January 2011 (R2 = 0.15–0.65 and SE = 0.017–0.038 cm3/cm3). The poor correlation on 10 January 2011 is mainly due to the continuous drought from the autumn of 2010 to the spring of 2011 in Nanjing, and the sensitivity of the TVDI to soil moisture is lower under water-stressed surfaces, with lower soil moisture in the root zone. Thus, the plots are relatively scattered, and this is reflected in the rather low values of R2. In the regions in which the land cover types are broadleaf mixed forest, Quercus acutissima and Pinus massoniana, the R2 between the TVDI and soil moisture on 18 May 2011 (R2 = 0.58–0.79) is higher than on 23 September 2011 (R2 = 0.46–0.66). These regions occupied 89.1% of the area of the Laoshan forest. The reason for this phenomenon is that the values of NDVI on 23 September 2011 are higher than those on 18 May 2011 (see Figure 8). The mean of NDVI on 23 September 2011 is 0.61, and the mean of NDVI on 18 May 2011 is 0.54. This can reduce the sensitivity of the TVDI to soil moisture under high vegetation cover [7].
Table 3. The coefficients e and f of the regression equation (see Equation (4)) between the TVDI and soil moisture, standard errors of soil moisture estimation (SE) and R2 values for linear fit of the data (units: cm3/cm3).
Table 3. The coefficients e and f of the regression equation (see Equation (4)) between the TVDI and soil moisture, standard errors of soil moisture estimation (SE) and R2 values for linear fit of the data (units: cm3/cm3).
DateLand Cover TypeefR2SE
10 January 2011Eucommia ulmoides–0.250.360.150.036
Quercus acutissima–0.240.360.540.030
broadleaf mixed forest–0.290.420.340.033
Cunninghamia lanceolata–0.310.430.540.024
coniferous and broadleaf mixed forest–0.250.400.280.038
Pinus massoniana–0.270.450.650.017
18 May 2011Eucommia ulmoides–0.290.390.380.037
Quercus acutissima–0.350.410.680.024
broadleaf mixed forest–0.300.410.580.031
Cunninghamia lanceolata–0.260.400.290.041
coniferous and broadleaf mixed forest–0.300.410.470.034
Pinus massoniana–0.380.470.790.023
23 September 2011Eucommia ulmoides–0.250.390.380.034
Quercus acutissima–0.290.420.660.026
broadleaf mixed forest–0.280.410.460.026
Cunninghamia lanceolata–0.290.440.360.023
coniferous and broadleaf mixed forest–0.450.480.800.015
Pinus massoniana–0.250.440.530.027
Earlier studies used TVDI to estimate soil moisture from different remote sensing data and produced different R2 values. These different remote sensing data are NOAA-AVHRR images (R2 = 0.23–0.81) [10,16] and Terra/Aqua MODIS images (R2 = 0.12–0.83) [7,8,17,18,19,20]. Gao et al. [21] used LANDSAT-5 TM images, but they did not calculate R2 values. For some studies, the R2 between the TVDI and soil moisture is still low [8,19]. The reason is that the TVDI values retrieved from remote sensing images suffer from scaling effects. For example, each pixel in the Landsat-5 TM image represents an area of 30 m × 30 m, but each observation point of in situ soil moisture measurement represents only a point on the soil surface, so the in situ measured soil moisture cannot ensure a perfect match with the corresponding pixel in the image.
Figure 8. Histograms of the NDVI for (a) 18 May 2011 and (b) 23 September 2011.
Figure 8. Histograms of the NDVI for (a) 18 May 2011 and (b) 23 September 2011.
Sustainability 07 11401 g008

4.4. Soil Moisture under Different Types of Land Cover

Table 3 shows that the standard errors of soil moisture estimation (SE) are low (SE = 0.015–0.041 cm3/cm3). Therefore, we have used the regression equation between the TVDI and soil moisture in Table 3 to calculate the soil moisture for the area associated with each type. Histograms for all six types are shown in Figure 9.
Figure 9. Histograms of the soil moisture under each type of land cover for the three scenes.
Figure 9. Histograms of the soil moisture under each type of land cover for the three scenes.
Sustainability 07 11401 g009aSustainability 07 11401 g009b
Figure 9 illustrates that the trends of soil moisture under each type of land cover for three different temporal patterns are similar to those of TVDI. The values of soil moisture on 23 September 2011 are higher than that on 18 May 2011, and the values of soil moisture on 18 May 2011 are higher than those on 10 January 2011. The mean values of soil moisture and variance under each type of land cover for the three scenes were calculated and are presented in Table 4.
Table 4 shows that the trends of soil moisture under the six types of land cover on 10 January 2011 are similar to those on 18 May 2011, and the mean values of soil moisture from low to high are listed in the following order: Eucommia ulmoides, Quercus acutissima, broadleaf mixed forest, Cunninghamia lanceolata, coniferous and broadleaf mixed forest and Pinus massoniana. However, on 23 September 2011, the mean value of soil moisture and variance under Eucommia ulmoides (mean = 0.2247, variance = 0.0014) is higher than that under Quercus acutissima (mean = 0.2198, variance = 0.0007). The reason for this phenomenon is that there was a small amount of cloud in the sky; although the region covered by cloud was screened out, the thin cloud also influences the accuracy of soil moisture estimation.
Table 4. The mean values of soil moisture and variance under each type of land cover for the three scenes (units: cm3/cm3).
Table 4. The mean values of soil moisture and variance under each type of land cover for the three scenes (units: cm3/cm3).
Types10 January 201118 May 201123 September 2011
MeanVarianceMeanVarianceMeanVariance
Eucommia ulmoides0.20720.00100.21490.00110.22470.0014
Quercus acutissima0.21300.00040.21670.00090.21980.0007
broadleaf mixed forest0.21540.00120.23600.00180.24480.0015
Cunninghamia lanceolata0.23220.00170.24080.00140.24540.0012
coniferous and broadleaf mixed forest0.25390.00320.27010.00220.28860.0016
Pinus massoniana0.27450.00090.28690.00070.29170.0006
According to Table 4 and Figure 9, the soil water content under the six types of land cover from low to high is listed in the following order: Eucommia ulmoides, Quercus acutissima, broadleaf mixed forest, Cunninghamia lanceolata, coniferous and broadleaf mixed forest and Pinus massoniana.

5. Conclusions

The present study used LANDSAT-5 TM images to explore the potential of TVDI in the Laoshan forest. The relationship between TVDI and volumetric soil moisture at the observation points is significant. Furthermore, this paper investigates the ability of the TVDI to capture temporal variations in surface moisture. The results are encouraging and show that the TVDI values on 10 January 2011 are higher than those on 18 May 2011 and 23 September 2011 because of the continuous drought in the winter of 2011. The TVDI values after precipitation are lower than for periods of no precipitation. Therefore, it is summarized that the TVDI, which is solely based on satellite observations, is feasible for monitoring soil moisture in the Laoshan forest, and it can reflect the soil moisture status for three different times of the year.
Over the last decade, local government and scientists have raised many concerns about the hydrological roles of forests, and most studies use forest resource survey data and mathematical simulations to estimate the spatial variation of water holding of the forest ecosystems [32]. However, forest resource survey data and mathematical simulations need a lot of in situ measurements and are cost-intensive. So far, substantial research has been carried out to retrieve soil moisture using remotely-sensed observations [6], but little research focused on investigating the differences of soil moisture under different tree species. In this paper, an effort is also made to investigate the soil moisture under the six types of land cover via the TVDI with the aid of three different temporal patterns, and the soil moisture from low to high is listed in the following order: Eucommia ulmoides, Quercus acutissima, broadleaf mixed forest, Cunninghamia lanceolata, coniferous and broadleaf mixed forest and Pinus massoniana.
However, cloud cover seriously influences surface temperature retrieval and often limits the applicability of the TVDI for estimating soil moisture at a regional scale. The effects of scaling also have an impact on the accuracy of remotely-sensed soil moisture. Additional work using meteorological data and hydrological models, such as the soil water under forests (SWUF) model [33], should be done to test and validate the robustness of the TVDI over large regions.

Acknowledgments

This work was supported by the NSFC (National Natural Science Foundation of China) International Partnership Project under Grant Number: 3136113034; NSFC (National Natural Science Foundation of China) Project under Grant Number: 41171324; the Platform Project of Ministry of Science and Technology of China under Grant Number: 2005DKA32306; NSFC (National Natural Science Foundation of China) Project under Grant Number: 41101315; NSFC (National Natural Science Foundation of China) Project under Grant Number: 71373125; and JSNSF (Jiangsu Natural Science Foundation of China) Project under Grant Number: BK20140980.

Author Contributions

Shulin Chen and Zuomin Wen designed the sites of in situ measurements and built the modeling. Hong Jiang and Xiuying Zhang helped to perform the experiments and analyzed the data. Qinjian Zhao analyzed the accuracy and reliability of the method. Yan Chen participated in the editing of the paper.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Seneviratne, S.I.; Corti, T.; Davin, E.L.; Hirschi, M.; Jaeger, E.B.; Lehner, I.; Orlowsky, B.; Teuling, A.J. Investigating soil moisture—Climate interactions in a changing climate: A review. Earth-Sci. Rev. 2010, 99, 125–161. [Google Scholar] [CrossRef]
  2. Mathys, A.; Coops, N.C.; Waring, R.H. Soil water availability effects on the distribution of 20 tree species in western North America. Forest Ecol. Manag. 2014, 313, 144–152. [Google Scholar] [CrossRef]
  3. Wu, Y.M. Challenges and countermeasures of water resources in the process of urbanization. Jiangsu Water Resour. 2003, 5, 29–30. (In Chinese) [Google Scholar]
  4. Skierucha, W.; Wilczek, A. A FDR sensor for measuring complex soil dielectric permittivity in the 10–500 MHz frequency range. Sensors 2010, 10, 3314–3329. [Google Scholar] [CrossRef] [PubMed]
  5. Skierucha, W.; Wilczek, A.; Szypłowska, A.; Sławiński, C.; Lamorski, K. A TDR-based soil moisture monitoring system with simultaneous measurement of soil temperature and electrical conductivity. Sensors 2012, 12, 13545–13566. [Google Scholar] [CrossRef] [PubMed]
  6. Grayson, R.B.; Western, A.W. Towards areal estimation of soil water content from point measurement: Time and space stability of mean response. J. Hydrol. 1998, 207, 68–82. [Google Scholar] [CrossRef]
  7. Patel, N.R.; Anapashsha, R.; Kumar, S.; Saha, S.K.; Dadhwal, V.K. Assessing potential of MODIS derived temperature/vegetation condition index (TVDI) to infer soil moisture status. Int. J. Remote Sens. 2009, 30, 23–39. [Google Scholar] [CrossRef]
  8. Chen, S.L.; Liu, Y.B.; Wen, Z.M. Satellite retrieval of soil moisture: An overview. Adv. Earth Sci. 2012, 27, 1192–1203. (In Chinese) [Google Scholar]
  9. Goetz, S.J. Multi-sensor analysis of NDVI, surface temperature and biophysical variables at a mixed grassland site. Int. J. Remote Sens. 1997, 18, 71–94. [Google Scholar] [CrossRef]
  10. Sandholt, I.; Rasmussen, K.; Andersen, J. A simple interpretation of the surface temperature vegetation index space for assessment of surface moisture status. Remote Sens. Environ. 2002, 79, 213–224. [Google Scholar] [CrossRef]
  11. Goward, S.N.; Xue, Y.K.; Czajkowski, K.P. Evaluating land surface moisture conditions from the remotely sensed temperature/vegetation index measurements: An exploration with the simplified simple biosphere model. Remote Sens. Environ. 2002, 79, 225–242. [Google Scholar] [CrossRef]
  12. Friedl, M.A.; Davis, F.W. Sources of variation in radiometric surface temperature over a tallgrass prairie. Remote Sens. Environ. 1994, 48, 1–17. [Google Scholar] [CrossRef]
  13. Carlson, T.N.; Gillies, R.R.; Perry, E.M. A method to make use of thermal infrared temperature and NDVI measurements to infer surface soil water content and fractional vegetation cover. Remote Sens. 1994, 9, 161–173. [Google Scholar] [CrossRef]
  14. Moran, M.S.; Clarke, T.R.; Inoue, Y.; Vidal, A. Estimating crop water deficit using the relation between surface–Air temperature and spectral vegetation index. Remote Sens. Environ. 1994, 49, 246–263. [Google Scholar] [CrossRef]
  15. Xin, J.F.; Tian, G.L.; Liu, Q.H.; Chen, L.F. Combining vegetation index and remotely sensed temperature for estimation of soil moisture in China. Int. J. Remote Sens. 2006, 27, 2071–2075. [Google Scholar] [CrossRef]
  16. Wang, C.Y.; Qi, S.H.; Niu, Z.; Wang, J.B. Evaluating soil moisture status in China using the temperature–vegetation dryness index (TVDI). Can. J. Remote Sens. 2004, 30, 671–679. [Google Scholar] [CrossRef]
  17. Holzman, M.E.; Rivas, R.; Piccolo, M.C. Estimating soil moisture and the relationship with crop yield using surface temperature and vegetation index. Int. J. Appl. Earth Obs. 2014, 28, 181–192. [Google Scholar] [CrossRef]
  18. Zhang, F.; Zhang, L.W.; Shi, J.J.; Huang, J.F. Soil moisture monitoring based on land surface temperature—vegetation index space derived from MODIS data. Pedosphere 2014, 24, 450–460. [Google Scholar] [CrossRef]
  19. Chen, J.; Wang, C.Z.; Jiang, H.; Mao, L.X.; Yu, Z.R. Estimating soil moisture using Temperature—Vegetation Dryness Index (TVDI) in the Huanghuai-hai (HHH) plain. Int. J. Remote Sens. 2011, 32, 1165–1177. [Google Scholar] [CrossRef]
  20. Rahimzadeh-bajgiran, P.; Omasa, K.; Shimizu, Y. Comparative evaluation of the Vegetation Dryness Index (VDI), the Temperature Vegetation Dryness Index (TVDI) and the improved TVDI (iTVDI) for water stress detection in semi-arid regions of Iran. ISPRS J. Photogramm. 2012, 68, 1–12. [Google Scholar] [CrossRef]
  21. Gao, Z.Q.; Gao, W.; Chang, N.B. Integrating temperature vegetation dryness index (TVDI) and regional water stress index (RWSI) for drought assessment with the aid of LANDSAT TM/ETM+ images. Int. J. Appl. Earth Obs. 2011, 13, 495–503. [Google Scholar] [CrossRef]
  22. Luan, H.J.; Tian, Q.J.; Yu, T.; Hu, X.L.; Huang, Y.; Du, L.T.; Zhao, L.M.; Wen, X.; Han, J.; Zhang, Z.W.; et al. Modeling continuous scaling of NDVI based on fractal theory. Spectrosc. Spect. Anal. 2013, 33, 1857–1862. (In Chinese) [Google Scholar]
  23. Sun, L.; Sun, R.; Li, X.; Liang, S.; Zhang, R. Monitoring surface soil moisture status based on remotely sensed surface temperature and vegetation index information. Agr. Forest Meteorol. 2012, 166, 175–187. [Google Scholar] [CrossRef]
  24. Mallick, K.; Bhattacharya, B.K.; Patel, N.K. Estimating volumetric surface moisture content for cropped soils using a soil wetness index based on surface temperature and NDVI. Agr. Forest Meteorol. 2009, 149, 1327–1342. [Google Scholar] [CrossRef]
  25. Chander, G.; Markham, B.L.; Helder, D.L. Summary of current radiometric calibration coefficients for Landsat MSS, TM, ETM+, and EO-1 ALI sensors. Remote Sens. Environ. 2009, 113, 893–903. [Google Scholar] [CrossRef]
  26. French, A.N.; Norman, J.M.; Anderson, M.C. A simple and fast atmospheric correction for spaceborne remote sensing of surface temperature. Remote Sens. Environ. 2003, 87, 326–333. [Google Scholar] [CrossRef]
  27. Qin, Z.; Karnieli, A.; Berliner, P. A mono-window algorithm for retrieving land surface temperature from Landsat TM data and its application to the Israel-Egypt border region. Int. J. Remote Sens. 2001, 22, 3719–3746. [Google Scholar] [CrossRef]
  28. Sobrino, J.A.; Jimnéez-muñoz, J.C.; Panolini, L. Land surface temperature retrieval from LANDSAT TM 5. Remote Sens. Environ. 2004, 90, 434–440. [Google Scholar] [CrossRef]
  29. Xiao, R.B.; Weng, Q.H.; Ouyang, Z.Y.; Li, W.F.; Schienke, E.W.; Zhang, Z.M. Land surface temperature variation and major factors in beijing, China. Photogramm. Eng. Rem. S. 2008, 74, 451–461. [Google Scholar] [CrossRef]
  30. Wang, F.; Qin, Z.H.; Song, C.Y.; Tu, L.L.; Karnieli, A.; Zhao, S.H. An improved mono-window algorithm for land surface temperature retrieval from landsat 8 thermal infrared sensor data. Remote Sens. 2015, 7, 4268–4289. [Google Scholar] [CrossRef]
  31. Duan, Y.; Wang, W.; Cai, X.J. Applied analyses on Palmer, SPEI and CI Indices of drought process in Yangtze-Huaihe River Basins during winter of 2010/spring of 2011. Plateau Meteorol. 2013, 32, 1126–1139. (In Chinese) [Google Scholar]
  32. Pollacco, J.A.P. A generally applicable pedotransfer function that estimates field capacity and permanent wilting point from soil texture and bulk density. Can. J. Soil Sci. 2008, 88, 761–774. [Google Scholar] [CrossRef]
  33. Paul, K.I.; Polglase, P.J.; O’Connell, A.M.; Carlyle, J.C.; Smethurst, P.J.; Khanna, P.K.; Worledge, D. Soil water under forests (SWUF): A model of water flow and soil water content under a range of forest types. Forest Ecol. Manag. 2003, 182, 195–211. [Google Scholar] [CrossRef]

Share and Cite

MDPI and ACS Style

Chen, S.; Wen, Z.; Jiang, H.; Zhao, Q.; Zhang, X.; Chen, Y. Temperature Vegetation Dryness Index Estimation of Soil Moisture under Different Tree Species. Sustainability 2015, 7, 11401-11417. https://doi.org/10.3390/su70911401

AMA Style

Chen S, Wen Z, Jiang H, Zhao Q, Zhang X, Chen Y. Temperature Vegetation Dryness Index Estimation of Soil Moisture under Different Tree Species. Sustainability. 2015; 7(9):11401-11417. https://doi.org/10.3390/su70911401

Chicago/Turabian Style

Chen, Shulin, Zuomin Wen, Hong Jiang, Qingjian Zhao, Xiuying Zhang, and Yan Chen. 2015. "Temperature Vegetation Dryness Index Estimation of Soil Moisture under Different Tree Species" Sustainability 7, no. 9: 11401-11417. https://doi.org/10.3390/su70911401

Article Metrics

Back to TopTop