Available via license: CC BY-NC 4.0
Content may be subject to copyright.
Drought Monitoring & Changes
in the Vegetation Cover of Aurangabad Region,
Maharashtra Over Last Two Decades:
A Comparative Study of Remote Sensing Indices
VCI and SVI Using Time Series MODIS Data
Product MOD13Q6
Rashmi Nitwane(B), Vaishali Bhagile, and Ratnadeep Deshmukh
CS&IT Department, Dr. B. A. M. University, Aurangabad, India
rashminitwane@gmail.com
Abstract. Aurangabad district is part of Marathwada region of Maharashtra,
India. The region suffers with frequent drought occurrences resulting in sub-
stantial crop yield losses. The times series analysis of the satellite imagery is
helpful in monitoring the short term and long-term changes and understand the
reasons for recurrent drought occurrences. In this study we analyze time series
multispectral data products obtained from MODIS the spectroradiometer on board
Terra. The dataset of MOD13Q6 (2001–2020) 250 m for the Aurangabad region
is investigated using remote sensing indices VCI and SVI. The statistical tools
and techniques were implemented to understand if there is any drought recurrence
pattern over the years and then time series regression analysis of both the VCI
and SVI indices were done to find out which remote sensing index gives better
results. After comparing the results of two indices it is confirmed that though
both indices gave same results for some statistical tests. The Generalized Linear
Regression analysis performance of SVI is better and can be used for predicting
the vegetation condition to some extent. Although the accuracy of such prediction
is low because of the unpredictability of the climatic variables and the changing
dynamics of the land use due to urbanization and human induced factors. This
study was also useful in understanding the limitations of Satellite imagery and
vegetation indices.
Keywords: multispectral ·vegetation condition index ·standard vegetation
index
1 Introduction
Drought monitoring and Assessment has become the need of an hour as droughts impact
the socio-economic condition of the country adversely. Droughts is one of the worst
natural calamities because it spreads slowly and impact the food security of the region.
© The Author(s) 2023
S. Tamane et al. (Eds.): ICAMIDA 2022, ACSR 105, pp. 260–275, 2023.
https://doi.org/10.2991/978-94-6463-136-4_24
Drought Monitoring & Changes in the Vegetation Cover 261
Though atmospheric and naturally occurring phenomena are very unpredictable but
Space technology has given us opportunities to observe and analyze this data and find
patterns to predict or forecast the occurrence of the event. Satellite based vegetation
indices has become an important method for examining agricultural droughts as no field
measurements, interpolation or large-scale modelling is required [16].
In recent years India has evolved from food deficient and import dependent nation to
a global agricultural powerhouse [1] still drought monitoring at regional scale in India
suffers a setback as Indian agriculture is characterized by agro-ecological diversities
in soil, rainfall, temperature, and cropping system [2]. Drought monitoring and devel-
opment of new methods for the regional level agricultural drought assessment can be
done through proper analysis of data products obtained from the satellites [3]. In the
current study we have taken Moderate Resolution Imaging Spectrometer (MODIS) time
series vegetation product and surface reflectance product to understand the changes in
vegetation dynamics of the region. MODIS is a key instrument aboard Terra and Aqua
Satellites. The MODIS instrument provides high radiometric sensitivity in 36 spectral
bands ranging in the wavelength from 0.4μm to 14.4 μm.[4].
The second section of this study elaborates the MODIS data products in general. The
version 006 product used in this study is planned for retirement soon, the reasons are
further elaborated in Sect. 2. The third section describes the study area, fourth section
summarizes the vegetation indices and material and methods used and, fifth section gives
the results and conclusions.
2 Moderate Resolution Imaging Spectrometer (MODIS)
MODIS Terra& Aqua are the key instruments on Earth Observation Satellite (EOS). Both
the instruments view the earth surface in 36 spectral bands and three spatial resolutions.
MODIS data products are useful for our understanding of the earth processes. The data
product used in this study is MODIS Terra product for vegetation cover. Out of 36
spectral bands Two bands are imaged at a nominal resolution of 250 m at nadir, with five
bands at 500 m, and the remaining 29 bands at 1 km. A ±55-degree scanning pattern
at the EOS orbit of 705 km achieves a 2,330-km swath and provides global coverage
every one to two days. Bands 1–19 are in nm whereas Bands 20–36 are in μm[4]. Table
1gives the general specifications of the Terra satellite [5].
Terra is now scheduled to exit the morning constellations of Earth science satellites
by October 2022 because of onboard fuel shortage [6]. Terra has five instruments on
board ASTER, CERES, MISR, MODIS and MOPITT and the terra’s orbital drift has
affected all the sensors of Terra though Impacts will be minimal, but will be noticeable
in data and imagery. Longer shadows will be visible in images of mountain landscapes.
The sensors views will become narrower leading to smaller swath widths as Terra drifts
to an earlier crossing time [7].
2.1 MODIS Data Products
Both MOD13Q1.006 is Land data product of MODIS. MOD13Q1 version 6 data are gen-
erated every 16 days at 250 m(m) spatial resolution as a level 3 product. MOD13Q1.006
262 R. Nitwane et al.
Table 1. MODIS Terra Specification
TERRA Specification
1. Orbit 705 km
10:30 AM
Descending mode
2. Scan rate 20.3 rpm, cross track
3. Swath Dimension 2330 km (cross track) by 10 km along track at Nadir
4. Telescope 17.78 cm diameter, off axis, afocal with intermediate field stop
5. Size 1.0 ×1.6 ×1.0 m
6. Weight 228.7 kg
7. Power 162.5 W (single orbit average)
8. Data Rate 10.6 Mbps(peak day time)
6.1 Mbps (orbital avg)
9. Quantization 12 its
10. Spatial Resolution 250 m (bands 1–2), 500 m (bands 3–7)
1000 m (bands 8–36)
11. Design Life 6 years
provides two vegetation layers first is Normalized Difference Vegetation Index (NDVI)
also referred as continuity index and the other one is Enhanced Vegetation Index (EVI)
which has improved sensitivity over high biomass regions. The algorithm chooses the
best available pixel values from the 16-day period [8]. In the current study we have
utilized the MOD13Q1.006 data set from period DOY 001_2001- DOY 001_2020.
It should be further noted that the collection 6 forward processing is planned to be
discontinued soon since there are issues detected in this product. The list of issues for
each product is summarized in Table 2.
The incorrect representations of the aerosol quantities in collection 6 MOD09 surface
reflectance products have impacted MOD13 vegetation index data products particularly
over arid bright surfaces [9].
3 Study Area
Aurangabad district is one of the 36 districts located in the state of Maharashtra, India.
It is located mainly in Godavari Basin and some of its part lays North West of Tapi
river basin. Total area of Aurangabad district is approx. 10.08 lakh hectares out of which
area under cultivation is 8.52 lakh hectares, 141.1 sq·km is urban and 99,587 sq·km
is rural area. The district occupies 1362 villages and is divided in 8 divisions which
are sub grouped into five subdivisions. The region is arid and the year is divided in
three seasons i.e., summer, Rainy (Monsoon) and winter. In Aurangabad rainy season
starts from month of June–September, winter season starts from October–February and
summer season starts from March- May. The average rainfall received is 734 mm and
Drought Monitoring & Changes in the Vegetation Cover 263
Table 2. MODIS Data Product issues
MOD13Q1.006 Vegetation Product Issues detected
1. Unexpected missing data in the last cycles of the year
2Incorrect instances of “No Data” and spikes in NDVI values
3. VI Usefulness Bits are not correctly assigned
MOD09A1 v006 Surface reflectance issue detected
4. The incorrect representations of the aerosol quantities in collection 6 MOD09
surface reflectance products have impacted MOD13 vegetation index data products
particularly over arid bright surfaces
Fig. 1. Map of Aurangabad, Maharashtra, India
min temp is as low as 5.6 °C and max temperature is 45.9 °C. There are two main crop
cycles followed Monsoon season crops are called Kharif crops and winter seasonal crops
are called Rabi. Major food grains grown in this region are Jowar, pearl millet, wheat
and gram, oilseeds soyabean is major crop and cotton is major cash crop of the region
[10]. Figure 1shows the map of the study region.
4 Vegetation Indices (VCI and SVI)
Remote sensing of vegetation is mainly performed by obtaining the electromagnetic
wave reflectance information from canopies using passive sensors [11]. MODIS Vegeta-
tion Indices product NDVI and EVI produced on 16-day intervals and at multiple spatial
resolutions provide consistent spatial and temporal comparisons of vegetation canopy
greenness, chlorophyll and canopy structure [12]. In this study MOD13Q1.006 250 m
264 R. Nitwane et al.
EVI vegetation product for the time period of 20 years is utilized The Enhanced Vege-
tation Index (EVI) minimizes the canopy soil variations and improves sensitivity over
dense vegetation conditions moreover EVI can be associated with stress and changes
related to drought [14].
Vegetation condition Index (VCI) focusses on the impact of drought stress on veg-
etation and can provide information on the onset, duration and severity of drought by
noting vegetation changes and comparing them with historical values [13]. The Eq. (1)
shows the calculation of VCI
VCIijk =EVIijk −EVIi,min
EVIimax −EVIimin ∗100 (1)
where,
EVIijk =The EVI value for the pixel i, during j time period for year k.
EVIi,min =The minimum value of the pixel i
EVIi,max =The maximum value of the pixel i
The other vegetation Index used in the current study is Standard Vegetation Index
(SVI). It is an estimate of the probability of occurrence of the present vegetation condi-
tion. SVI is based on the calculations of z-score of each pixel. The z-score is a deviation
from mean in units of standard deviation calculated from the EVI values for each pixel
location [15]. The Eq. (2) describes SVI equation.
Z_ijk =(EVI_ijk −EVImean_ij)/(σ _ij)(2)
where,
EVIijk =The EVI value for the pixel i, during j time period for year k
EVImean ij =Mean EVI value for the pixel i during the j time period
σ_ij =Standard deviation of pixel i during j time period
In this study both VCI and SVI are used in conjunction with the EVI times series data
product of MOD13Q1.006 from 2001–2020 for the assessment of vegetation condition
in Aurangabad region of Maharashtra. The aim of this study is to compare and identify
the strengths and weakness of both in the assessment of agricultural drought on a regional
scale.
4.1 Material and Methods
Input Data
The present study utilizes 250 m resolution Terra MODIS Enhanced Vegetation Index
(EVI) time series starting from year DOY 001_2001 to DOY001_2020. Sixteen-day
composites of EVI and the corresponding pixel reliability layers were extracted from
the MODIS product MOD13Q1, version6 resulting in 23 Days of the year (DOY) per
year. In total 460 images were analyzed to assess the agricultural drought conditions
using VCI and SVI.
Drought Monitoring & Changes in the Vegetation Cover 265
Data Processing and Analysis
The MODIS data shows data gaps towards the later part of the year therefore the analysis
of vegetation condition over the Aurangabad region is based on the assessment of the
deviations of MODIS 16-day EVI measurements therefore EVI mean, median and stan-
dard deviation was calculated for the period of 19 years i.e., 2001–2020. The Skewness
in the data set was observed using the Karl Pearson method. Similarly, mean, median and
standard deviation of Standard Vegetation Index (SVI) and Vegetation Condition Index
(VCI) were calculated and compared. We conducted non parametric statistical tests on
the mean values of VCI and SVI to detect the trend in the data. The list of statistical tests
is shown in Table 4. A brief description of the statistical functions is given in Table 5;
the results are discussed in Sect. 5.
To develop a prediction model on VCI and SVI firstly pixel wise point wise values of
the study area was extracted of the time period 2001–2020, then splitting the data frame
in to train and test set we then calculated two sets of row quantiles with probabilities
0.50 and 0.75 was calculated on both the VCI and SVI train data frames. The flowchart
of the methodology is shown below.
Non-parametric Tests
Non parametric trend test is more suitable for natural time series data set because assump-
tions required for parametric testing are not usually present in this type of data set [17].
To identify whether or not there is a trend in the dataset Cox and Stuart trend test is
implemented. The idea of the Cox-Stuart test is based on the comparison of the first and
the second half of the sample [18]. The statistical hypotheses in testing for the trend in
the series of random variables are
H0 =No monotonic trend exists in the series
Ha =Monotonic trend exist in the series
Mann Kendall trend test is considered to be appropriate to identify climate change
and trend identification [19]. The correlation can significantly influence the results in
positive correlation there is a possibility of rejecting the null hypothesis whereas in
negative correlation acts to accept the null hypothesis [20]. The other rank based non
parametric test called as Pettit’s test for detection of abrupt change in the time series
[21]. This test gives the point of change in the time series dataset. Additionally, Wallis
and Moore phase frequency test was conducted to verify the significance of the trend
and the Sen’s slope test was conducted to determine the magnitude of the trend.
Regression Analysis
Regression analysis is most commonly used statistical tool to analyze the time series data
set. The three distinguished processes in regression analysis are (i) model selection (ii)
Parameter estimation and (iii) prediction of future values [22]. Regression analysis helps
us to understand the relationship between the variables as well as it can be used for the
purpose of prediction. Till date only Ordinary Linear square regression (OLS) method is
implemented in spatial time series data [24]. Therefore, out of various regression tech-
niques we have used the Generalized linear regression to determine whether prediction
modeling of the Remote Sensing Indices Standard Vegetation Index (SVI) and Vegeta-
tion Condition Index (VCI) is possible or not. The idea is to use a general exponential
family for the response distribution. In addition to real and binary responses, GLMs can
266 R. Nitwane et al.
Fig. 2. Flowchart for regression modelling
handle categorical, positive real, positive integer, and ordinal responses. GLM help us
build probabilistic predictors of many kinds of responses. GLM [23]. The flowchart in
Fig. 2. Explains the steps followed for modelling Generalized Linear regression model.
And Table 5shows the results of generalized linear regression analysis on the extracted
point wise values of the time series data set. The Data processing and statistical analysis
is done using R studio.
5 Results and Discussion
5.1 Observed Changes in MODIS Vegetation Product MOD13Q1.006 EVI Data
Set
To observe the spatial skewness in the vegetation product of MODIS the EVI mean,
EVI Median is calculated for the time period 2001–2020. It is observed that though
the impact is minimal but it is observable at regional scale and can be considered as
one of the indicators of vegetation shift. The EVI Skewness for the region over the
time period 2001- 2020 was equal to 0.5 which implicates that vegetation paradigm
shift is occurring slowly in the region. Comparing Differences in standard deviation and
skewness of remote sensing indices SVI and VCI (Fig. 3).
5.2 To Conduct Performance Analysis of the Two Remote Sensing Indices
We extracted mean, median of both the indices for the period of two decades [2001–
2020]. It is observed that there is very little difference in standard deviation of SVI values
Drought Monitoring & Changes in the Vegetation Cover 267
Fig. 3. (a) Mean EVI from 2001–2020. (b) EVI Median from 2001–2020. (c) Skewness of EVI
from 2001–2020
over the years but VCI Standard deviation has visible fluctuations over the years. After
finding the skewness over the years for both the indices the results yield that though
both indicated that years 2011 and 2014 are negatively skewed and is an indication of
deviation from normal and of severe drought stress. SVI was also able to detect negative
skewness for the years 2010, 2017 and 2018 clearly. Table 3shows the Mean, Median
Standard Deviation and skewness of both SVI and VCI and the comparison chart is
displayed in Fig. 4(a) and Fig. 4(b).
5.3 Statistical Results of VCI_Mean and SVI_Mean
The statistical tests performed on VCI and SVI mean values gave us following
interpretations
1. There is Monotonic trend in the data set (observed by MK test and Cox & Stuart
test)
2. Probable change point at time k is of period 9 (Pettit’s test)
3. The series is significantly different from randomness (Wallis and Moore test)
4. The positive value of Tau means there is increasing trend. The tau values vary from +
1to−1 in this case tau is 0.33 for both the VCI mean and SVI mean which signifies
slightly increasing trend
5. Sen slope which estimates overall slope of the time series and is considered to be
test of magnitude of the time series shows higher magnitude of the change for VCI
(1.44) than SVI (0.0519)
268 R. Nitwane et al.
Table 3. Mean, Median, Deviation and Skewness of SVI & VCI
Year SVl_Mean SVl_Median SVl_Stddev Skewness _SVI VCl_Mean VCl_Median VCl_Stddev Skewness_VCI
2001 −0.557842 −0.682234 0.831614 0.448738 29.285310 24.016640 25.325420 0.624117
2002 −0.441379 −0.536129 0.803442 0.353792 32.345350 28.495140 23.792990 0.485464
2003 −0.564494 −0.693425 0.752195 0.514220 29.118640 24.618200 23.825380 0.566678
2004 −0.458086 −0.562361 0.730806 0.428053 31.904130 28.042640 22.431370 0.516440
2005 −0.222535 −0.254647 0.617362 0.156043 38.353720 36.899770 20.237160 0.215535
2006 −0.047574 −0.069237 0.627305 0.103597 43.134730 41.873500 20.280120 0.186572
2007 −0.068700 −0.082897 0.697900 0.061026 42.651160 41.779350 21.574870 0.121225
2008 −0.175545 −0.218438 0.628758 0.204654 39.718270 37.879860 19.992060 0.275872
2009 −0.459463 −0.500380 0.621450 0.197525 31.941560 30.538240 19.689110 0.213821
2010 1.031398 0.972434 0.777391 0.227544 72.483090 72.756060 20.729730 −0.039504
2011 0.806118 0.861679 0.826416 −0.201693 66.586340 68.258770 23.801980 −0.210793
2012 −0.107332 −0.129613 0.629794 0.106134 41.665240 40.468910 20.078560 0.178747
2013 −1.013263 −1.072020 0.716493 0.246022 17.136140 11.828040 18.878440 0.843517
2014 1.268227 1.322712 0.785615 −0.208061 79.004970 84.034170 21.508850 −0.701460
2015 −0.037956 −0.079485 0.842884 0.147811 43.575820 41.933500 24.348760 0.202350
2016 −0.099889 −0.152975 0.784704 0.202955 41.997060 40.121950 23.548290 0.238885
2017 0.530887 0.560125 0.831134 −0.105538 59.185200 59.145430 25.327350 0.004711
2018 0.566413 0.579991 0.836356 −0.048705 60.185170 60.217240 24.246890 −0.003968
2019 −0.807389 −0.881798 0.735128 0.303656 22.556210 18.827050 20.008870 0.559127
2020 0.858476 0.859771 0.863333 −0.004502 68.018250 68.010870 22.538380 0.000983
Drought Monitoring & Changes in the Vegetation Cover 269
0
5
10
15
20
25
30
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
Standard Deviation
SVI_Stddev VCI_Stddev
-1
-0.5
0
0.5
1
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
Skewness _SVI Skewness_VCI
(a)
(b)
Fig. 4. (a) Comparison of VCI and SVI by Standard Deviation. (b) Skewness of VCI vs SVI
6. Low p values signify that the test reject the null hypothesis and is statistically
significant.
270 R. Nitwane et al.
Table 4. Statistical Results [17,18]
Sr.
No.
Name of the test Result VCI Result SVI
1. Cox and Stuart Trend
test
data: vci_mean
z=1.6783, n =20, p-value
=0.09329
alternative hypothesis:
monotonic trend
data: svi_mean
z=1.6783, n =20, p-value
=0.09329
alternative hypothesis:
monotonic trend
2. Mann-Kendall trend test data: vci_mean
z=2.1089, n =20, p-value
=0.03496
alternative hypothesis: true S
is not equal to 0
sample estimates:
SvarStau
66.0000000 950.0000000
0.3473684
data: svi_mean
z=2.044, n =20, p-value =
0.04095
alternative hypothesis: true S
is not equal to 0
sample estimates:
SvarStau
64.0000000 950.0000000
0.3368421
3. Pettitt’s test for single
change-point detection
data: vci_mean
U* =55, p-value =0.2305
alternative hypothesis:
two.sided
sample estimates:
probable change point at
time K =9
data: svi_mean
U* =55, p-value =0.2305
alternative hypothesis:
two.sided
sample estimates:
probable change point at
time K =9
4. Wallis and Moore
Phase-Frequency test
data: vci_mean
z=0.83419, p-value =
0.4042
alternative hypothesis: The
series is significantly
different from randomness
data: svi_mean
z=0.83419, p-value =
0.4042
alternative hypothesis: The
series is significantly
different from randomness
5. Sen’s slope data: vci_mean
z=2.1089, n =20, p-value
=0.03496
alternative hypothesis: true z
is not equal to 0
95 percent confidence
interval:
0.08295624 2.26710065
sample estimates:
Sen’s slope 1.447817
data: svi_mean
z=2.044, n =20, p-value =
0.04095
alternative hypothesis: true z
is not equal to 0
95 percent confidence
interval:
0.001860815 0.083704114
sample estimates:
Sen’s slope 0.05197757
Drought Monitoring & Changes in the Vegetation Cover 271
5.4 Regression Analysis Result of SVI and VCI
Regression analysis after extraction of year wise per pixel SVI and VCI values from the
raster stack it is observed that Generalized Linear Model (GLM) performs better in case
of SVI. The results are summarized in Table 5.
Table 5. Regression Model Summary for SVI &VCI [23,24]
Regression
Models
Summary of Results
Generative
Linear Model
Call:SVI
glm(formula = formula1 ~ formula2, data = traindf, method =
"glm.fit")
Deviance Residuals:
Min 1Q Median 3Q Max
-1.03362 -0.14694 0.01296 0.15875 0.74595
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.397181 0.001836 -216.3 <2e-16 ***
formula2 0.619747 0.003866 160.3 <2e-16 ***
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
(Dispersion parameter for gaussian family taken to be 0.05167335)
Null deviance: 2966.5 on 31715 degrees of freedom
Residual deviance: 1638.8 on 31714 degrees of freedom
AIC: -3958.5
Number of Fisher Scoring iterations: 2
Call:VCI
glm(formula = formula1 ~ formula2, data = traindf, method =
"glm.fit")
Deviance Residuals:
Min 1Q Median 3Q Max
-45.107 -4.194 0.835 5.099 18.296
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -5.151033 0.151186 -34.07 <2e-16 ***
formula2 0.827144 0.002709 305.31 <2e-16 ***
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
(Dispersion parameter for gaussian family taken to be 51.28153)
Null deviance: 6406598 on 31715 degrees of freedom
Residual deviance: 1626342 on 31714 degrees of freedom
AIC: 214886
Number of Fisher Scoring iterations: 2
272 R. Nitwane et al.
6 Conclusion
The objective of this research was to understand the scope and challenges in the devel-
opment of prediction system of remote sensing indices. In the process of our study we
found that due to changing land use land forms either occurring due to natural causes
or by human induced changes there is impact on these remote sensing indices. Some of
the key points of this study are listed below.
•MOD13Q1.006 EVI data set contained unexpected missing data values after Julian
Day of the Year (DOY) 177 for all years.
•The Remote sensing vegetation indices SVI and VCI extracts the vegetation informa-
tion successfully though both performs statistically equivalent the magnitude analysis
(Sen’s slope) shows that there is higher change in VCI than SVI.
•There is less skewness observed in SVI than in VCI
•The strength of these indices is that they give real time information about the vegetation
condition at regional scale and real time monitoring is possible but use of these indices
for prediction purpose with greater accuracy needs more robust models and systems.
The major challenges observed in using Remote sensing indices for prediction includes
missing values in the raw data and spatial skewness (Tables 6and 7).
Table 6. Result Summary GLM for SVI
Result Summary
Generalized Linear Model SVI 2001-2020 Output
1. Predicted values
2. Theoretical
Quantiles
3. Residuals vs
Leverage
Drought Monitoring & Changes in the Vegetation Cover 273
Table 7. Result Summary GLM for VCI
1. Predicted val-
ues
2. Theoretical
Quantiles
3. Residuals vs
Leverage
Result Summary
Generalized Linear
Model
VCI 2001-2020 Output
References
1. Indian Agriculture: Achievements and Challenges by Atri Mukherjee, Priyanka Bajaj,
Rishabh Kumar and Jobin Sebastian of the Department of Economic and Policy Research,
Reserve Bank of India.in RBI Bulletin January 2022.
2. INDIAN AGRICULTURE – AN INTRODUCTION Submitted to Fourth Session of the
Technical Committee of APCAEM 10–12 February 2009, Chiang Rai, Thailand.
3. Thenkabail, P. S.; Gamage, M. S. D. N.; Smakhtin, V. U. 2004. Theuse of remote sensing data
for drought assessment and monitoring in Southwest Asia. Research Report 85. Colombo,
Sri Lanka: International Water Management Institute.
4. https://modis.gsfc.nasa.gov/about/
5. https://modis.gsfc.nasa.gov/about/specifications.php
6. https://modis-land.gsfc.nasa.gov/news.html
7. https://terra.nasa.gov/about/terra-orbital-drift-information/impacts-from-orbital-drift
8. Didan, K. (2015). MOD13Q1 MODIS/Terra Vegetation Indices 16-Day L3 Global 250m SIN
Grid V006. NASA EOSDIS Land Processes DAAC. Accessed 2022-08-24 from https://doi.
org/10.5067/MODIS/MOD13Q1.006
9. Vermote, E. (2015). MYD09A1 MODIS/Aqua Surface Reflectance 8-Day L3 Global 500m
SIN Grid V006 . NASA EOSDIS Land Processes DAAC. Accessed 2022-08-24 from https://
doi.org/10.5067/MODIS/MYD09A1.006
274 R. Nitwane et al.
10. https://aurangabad.gov.in/en/district-profile/
11. Jinru Xue, Baofeng Su, “Significant Remote Sensing Vegetation Indices: A Review of Devel-
opments and Applications”, Journal of Sensors, vol. 2017, Article ID 1353691, 17 pages,
2017. https://doi.org/10.1155/2017/1353691
12. https://modis.gsfc.nasa.gov/data/dataprod/mod13.php
13. Kogan, F.N., 1995: Application of vegetation index and brightness temperature for drought
detection. Advances in Space Research, 15(11): 91–100. https://doi.org/10.1016/0273-117
7(95)00079-T.
14. Huete, A., K. Didan, T. Miura, E.P. Rodriguez, X. Gao and L.G. Ferreira, 2002: Overview
of the radiometric and biophysical performance of the MODIS vegetation indices. Remote
Sensing of Environment, 83(1):195–213.
15. Peters, A.J., Walter-Shea, E.A., Ji, L., Viña, A., Hayes, M.J., & Svoboda, M.D. (2002).
Drought Monitoring with NDVI-Based Standardized Vegetation Index. Photogrammetric
Engineering and Remote Sensing, 68, 71–75.
16. Reinermann S, Gessner U, Asam S, Kuenzer C, Dech S. The Effect of Droughts on Vegetation
Condition in Germany: An Analysis Based on Two Decades of Satellite Earth Observation
Time Series and Crop Yield Statistics. Remote Sensing. 2019; 11(15):1783. https://doi.org/
10.3390/rs11151783
17. Hirsch, R. M., and Slack, J. R. (1984), A Nonparametric Trend Test for Seasonal Data With
Serial Dependence, Water Resour. Res., 20(6), 727–732, https://doi.org/10.1029/WR020i006
p00727 .
18. D. R. COX, A. STUART, SOME QUICK SIGN TESTS FOR TREND IN LOCATION AND
DISPERSION, Biometrika, Volume 42, Issue 1–2, June 1955, Pages 80–95, https://doi.org/
10.1093/biomet/42.1-2.80
19. Wang F, Shao W, Yu H, Kan G, He X,Zhang D, Ren M and Wang G (2020) Re-evaluation of
the Power of the Mann-Kendall Test for Detecting Monotonic Trends in Hydrometeorological
Time Series.Front. Earth Sci. 8:14. https://doi.org/10.3389/feart.2020.00014
20. Berger, Labeyrie, L. D., Berger, W. H., Wolfgang H., & Labeyrie, L. D., Laurent D. (1987).
Abrupt climatic change : evidence and implications/edited by W.H. Berger and L.D. Labeyrie
21. Gharechaee, Hamidreza & Sajedi Hosseini, Farzaneh & Khosravi, Hassan & Nasrollahi,
Mohammad. (2016). Trend analysis and detection of precipitation fluctuations in arid and
semi-arid regions.
22. Geoffrey H. Moore & W. Allen Wallis (1943) Time Series Significance Tests Based on Signs
of Differences, Journal of the American Statistical Association, 38:222, 153–164, https://doi.
org/10.1080/01621459.1943.10501791
23. Lehtonen, Risto & Särndal, Carl-Erik & Veijanen, Ari. (1999). Generalized regression and
model-calibration estimation for domains: Accuracy comparison.
24. Walberg, H.J. (1971). Generalized Regression Models in Educational Research.
Drought Monitoring & Changes in the Vegetation Cover 275
Open Access This chapter is licensed under the terms of the Creative Commons Attribution-
NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/),
which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any
medium or format, as long as you give appropriate credit to the original author(s) and the source,
provide a link to the Creative Commons license and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter’s Creative
Commons license, unless indicated otherwise in a credit line to the material. If material is not
included in the chapter’s Creative Commons license and your intended use is not permitted by
statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder.
