NDVI, Groundwater, Remote Sensing, Green leaf, TNDVI, VI, Digital Image Processing techniques, Ground Water, Normalized Difference Vegetation Index, Vegetation Health as a Groundwater Indicato, Basic principle of NDVI, Near Infrared Band value,

Normalized Difference Vegetation Index (NDVI) Analysis: Use of Vegetation Health as a Groundwater Indicator


Description of NDVI Technique to understand the health of vegetation as Groundwater indicator with GIS as a tool for analysis. A Case Study of Cuddapah District in Andhra Pradesh (India)


Arka Prava Mukherjee & Dhiraj Mohan Banerjee

Normalized Difference Vegetation Index (NDVI) Analysis, NDVI, Groundwater, Remote Sensing, Green leaf, TNDVI, VI, Digital Image Processing techniques, Ground Water, Normalized Difference Vegetation Index, Vegetation Health as a Groundwater Indicato, Basic principle of NDVI, Near Infrared Band value,

Updated On: 12/22/2005

Normalized Difference Vegetation Index (NDVI) Analysis, NDVI, Groundwater, Remote Sensing, Green leaf, TNDVI, VI, Digital Image Processing techniques, Ground Water, Normalized Difference Vegetation Index, Vegetation Health as a Groundwater Indicato, Basic principle of NDVI, Near Infrared Band value,
NDVI, Groundwater, Remote Sensing, Green leaf, TNDVI, VI, Digital Image Processing techniques, Ground Water, Normalized Difference Vegetation Index, Vegetation Health as a Groundwater Indicato, Basic principle of NDVI, Near Infrared Band value,

With the advent of satellite remote sensing it has become possible to understand the green leaf concentration or chlorophyll status of vegetation for a large area of the earth surface with the help of a single digital image. Out of the numerous Digital Image Processing techniques (like TNDVI, VI etc.) used; NDVI (or Normalized Difference Vegetation Index) happens to be the most widely used technique to help understand the vegetation health status [1-3]. This technique not only highlights the vegetated areas of an image but also gives an idea regarding as to how healthy the plants are.

The basic equation behind this operation can be expressed as: 

NDVI = (NIR – R) / (NIR + R)

where , NIR = Near Infrared Band value , R = Red Band value, recorded by the satellite sensor. This equation can be better understood from Figure 1.

If one looks at Figure 1(a) one can see that, the plant/vegetation, which is healthy, absorbs more Red radiation as compared to the stressed vegetation shown in Figure 1(B). Thus when one computes the NDVI equation in real values (i.e. 8 % becomes 0.08 and so on), a number is obtained representing the health status of the plant involved ¾ for example , 0.72 for the healthy plant, as shown in Figure 1(A). An important fact of the obtained values is that for a given NDVI image the resulting NDVI pixel value always ranges from -1 to +1. Also it is important to note that the areas devoid of any vegetation give a negative value or a value close to zero. In simpler words a –ve number or a number close to zero means no vegetation and a number close to +1 (0.8 - 0.7) represents Luxurious vegetation.

Therefore, when an array of such NDVI numbers is computed and projected in the form of an image; a grey scale image is obtained where the grey scale intensity of each pixel is proportional to NDVI value it represents (Figure 2).

Though, the health of a plant depends on several environmental factors ¾ it is often found that, for a large area the vegetation health largely depends on; as to how much moisture is available to the root zone of the plants. And it is this property that is exploited as an indirect indicator of the availability of groundwater below the surface of the earth. 

Thus during the dry spell of the year (i.e. April to June) a good quality vegetation in an area is often found to be associated with shallow ground water levels. Keeping this fact in view, anomalous zones of luxurious and healthy vegetation over an area are often marked by explorer as favourable targets for further detailed field surveys. Thus a NDVI image not only saves the time and energy but also brings down the cost of otherwise extensive field surveys. 

Figure 2 shows the NDVI image of a part of the Mugamureru River basin of Cuddapah district, Andhra Pradesh, India. The satellite image taken by IRS – 1D LISS III sensor, represents the vegetation condition as on 23rd March 2001. In this image, the brighter white areas represent areas with luxurious/ healthy vegetation as compared to the dull white areas showing areas where vegetation is stressed (refer the grey scale intensity bar in Figure 2). The dark areas represent areas devoid of any vegetation.

Figure 3 is the pseudocolour representation of the same image. In this image the intensity values are classified by Density, Slicing the pixels values into four classes namely, Luxurious vegetation, Healthy vegetation, Stressed vegetation and areas with No vegetation. It can be observed that the areas close to the dry river channels have a good concentration of Luxurious and healthy vegetation as compared to rest of the area. This is primarily due to better groundwater supply and moisture replenishment from the alluvial deposits of the rivers. As one moves away from the stream the density of the luxurious / healthy vegetation rapidly decreases. However small specks of Luxurious and healthy vegetation can be still observed for example the area marked as ‘A’ as shown in Figure.3.

Thus, apart from the vicinity of the riverbeds, the interior areas that show signs of luxurious/healthy vegetation are considered favorable targets for the next phase of a detailed groundwater survey. However as an essential practice the results obtained from an NDVI analysis are usually matched with other outputs of groundwater modeling themes before finally outlining the groundwater potential zones. 

Acknowledgement
APM & DMB are thankful to UGC for providing the funds.


References
1. Campbell, J.B., Introduction to Remote Sensing, 2nd Ed. 1996, The Guilford Press, NY, p444 – 479.

2. Gupta, R.P. Remote Sensing Geology, 1st Ed. 1991, Springer-Verlag, NY, p32-33.

3. Su. Z. Remote Sensing of land use and Vegetation for mesoscale hydrological studies (2000) Int. J. Remote Sensing, 21(2), 213-233.


 
Back to top