|
Abstract
GreenRevolution Gujarat, India SatelliteImagery
Dams Irrigation Desertification
|
Part II: How can vegetation change data be used to make conclusions about land use change in this region over this period?
Now that we have derived quantitative measures of vegetation in the region, we can begin to look at the implications of these vegetation changes for large-scale land use and land cover change patterns. I have separated out the major vegetation change trends that are shown in the data. These include the direct submergence and destruction of agricultural and forested land by large-scale dams, as well as dramatic increases in agriculture in the region due to new access to irrigation. Additionally I have used the additional data provided by SMA to analyze the use of this data for determining the extent and patterns of desertification and forest thinning in particular areas over the 1970’s.
The major pattern of dramatic vegetation decrease in both images can be easily attributed to the construction and resulting reservoirs of large-scale dams. Estimates from the SMA vegetation change results show that the Ukai dam alone drowned 128.02 km2 of vegetated areas over the eight-year period.[12] This is a significant disruption of the landscape and only reiterates the controversial nature of dams for the region. (See Figure 22) Many studies have analyzed the environmental impact of dams in India using remote sensing. For example, R. Nagarajan (2000) conducted a multi-temporal analysis of the Dudhganga Dam by overlaying submerged areas with geocoded data on village locations and infrastructure maps (roads, mines/mineral deposits, archaeological and historical monuments, agricultural areas, forest reserves, and land ownership boundaries) to determine the impact on the land due to direct submergence, as well as less-direct degradation.
Background
on dams in the region Dams are currently the most controversial environmental issue in India. Dams submerge natural habitats, displace hundreds of thousands of people, increase waterborne disease (Krishna 1996), and have even been blamed for minor earthquakes like the Konya dam tremors in 1987 (Tapas,1993). Habitat loss can be extreme depending on the size of the reservoir and the local environment. In Karnataka from 1956-1980, 18% (41,068 ha) of the total forest area lost was directly attributed to hydroelectric projects according to government sources (Tapas,1993). It has been estimated that by 1985 dams in India had drowned half a million hectares of forest, or approximately one tenth of the area that had benefited from canal irrigation (Agarwal & Narain, 1985). Unique habitat loss and widespread displacement of villages has caused a great amount of controversy around many dams from the 1960’s through today. In 1983 one project, called Silent Valley, was suspended because uninhabited, biologically rich tropical forests were planned to be submerged as a result (Krishna 1996). Today the Narmada Basin Development Program, planning 31 large dams and displacement of over one million people, remains at the forefront of politically charged issues in India (Agarwal & Narain, 1985). Additionally, in this particular region, a large amount of the areas to be submerged are currently farmed with cash crops, causing a large financial burden for local inhabitants (Krishna 1996). Dams are also extremely beneficial to much of the population in the region, providing electricity, drinking water, and irrigation to millions in an area where it is especially important to curb the impacts of cyclical droughts. Additionally, the water crisis in this region has been exasterbated by the loss of forest and grasslands, leaving soil bare and unable to retain water gained through precipitation (SAI, 1992). Many planners and politicians see large dams as the most efficient way to solve the irrigation and energy problems of semi-arid regions in developing countries (Mehta, 1997). For this reason the Indian government, with the help of foreign investment banks, has invested huge amounts of money in the construction of dams on most rivers in the region. The biggest dams in the country are located in eastern Gujarat and many more are still being planned today. One of the largest and most controversial is the Ukai dam, (see Figure 22). It began construction in 1971 and included as net catchment area of 62,225 km2. Unfortunately performance was slowed by siltation. The engineers assumed a siltation rate of 1.47 hectometers/100 km2 of catchment, however the actual rate was ten times that, 14.29 hectometers/100 km2 (Tapas, 1993). Additionally benefits were drastically varied amongst the population. Only 3,500 of the 18,500 families displaced by the Ukai dam were resettled (Agarwal & Narain, 1985). Furthermore, emphasis of water reserve tanks were generally towards irrigation and not drinking water, exasterbating already strained relations between the landed and landless rural populations (Mehta, 1997). Overall, dam projects have remained controversial around issues of technological strategies, environmental impacts, distribution of benefits, resettlement, foreign assistance, and human rights (SAI, 1992).
Irrigation: who is benefiting, large-scale vs. villages Interestingly, the major increases in vegetation cover in the region can also be attributed to dams and improved irrigation in the region. Figure 23 shows areas of huge increases in vegetation cover (from 95-419%). In order to get a better sense of the land uses in this area I supplemented my data with high spatial resolution images, such as the data acquired by the Landsat 7 satellite shown in Figure 24. This imagery was useful in elucidating ground features, typical plot size, and seasonal land uses, despite being acquired more than twenty years after the Landsat MSS data. From the detail supplied by the high resolution data is clear that the areas of drastic vegetation increase from 1976-1980 seem to correspond to large agricultural land holdings or cooperatives that irrigate and therefore harvest simultaneously. However, these benefits due to extensified irrigation on agriculture, are extremely localized and unevenly distributed. When specific known villages were compared to the areas pictured above this was clearly illustrated. The estimated vegetation increase for the villages pictured in Figure 6 averaged at 44%, while these large-scale plots had average increases of 242%.
Background on distribution of benefits of irrigation, villages Irrigation in the region has historically been unequally distributed and inefficiently used. Distinctions between classes and geographic areas were and remain sharply contrasting. As development continues, upper income groups show an increase in per capita consumption expenditure, the lower groups remain stagnant, while the bottom 5% show a decline. On the Village level, electrification, facilities of public water supply, extensive canal irrigation, and the existence of a few private telephone connections, remains non-existent in many rural India villages (Ansari, 1991). Agricultural land in India is unequally distributed, with the top 22% of the rural population owning 57% of the land and the bottom 24% owning only 3.7% in 1991, although this is an improvement from earlier decades (SAI, 1992). Large-scale dam irrigation projects generally have only 30-35% efficiency in this particular region due to seepage in unlined canal systems (SAI, 1992; Agarwal & Narain, 1985). A report by the Central Water and Power Commission in 1967 revealed that 71% of the water is lost in transit from the reservoir to the field (Agarwal & Narain, 1985). This inefficient irrigation can in turn cause further land degradation. “From the experience of major and medium irrigation works in India, it is evident that its benefits in arid areas though spectacular for the first 10-20 years, gradually get reduced and a considerable portion of the land gets deteriorated because of waterlogging and salinity.” DR Bhumbla, former Agriculture Commissioner (Agarwal & Narain, 1985). Water tables rise as a result of dams and irrigation, and this can have the effect of waterlogging and salinization in fields situated above already high water tables (Agarwal & Narain, 1985).
Desertification is also a major problem associated with semi-arid environments during the Green Revolution. Because sand fraction data was generated for each pixel in the process of vegetation change analysis through SMA, these data sets can be used to evaluate the extent to which desertification occurred in the region. The two images shown in Figures 25 & 26 illustrate the soil fraction change from 1972-1976 and then 1976-1980, classified according to magnitude of increase. As you can see, there seemed to be a serious increase in soil/sand fraction cover over the first four-year period and then a rebound of many areas over the next four years. As the precipitation data indicates (see Figure 27), this may very well be part of a drought signal, as there was significantly less over all water in the first period than over the second. In 1972 there was a major drought in the region forcing farmers to migrate into canal irrigated regions or nearby towns to receive aid, as wells in towns dried up completely (Ansari, 1991). When these changes are aggregated over the eight years there is clearly a large increase in sand cover in the image (see Figure 28). From 1972 to 1980, bare sandy soil cover increased 10% or more over 305.67 km2, or 14.5% of the area shown in the image. This clearly indicates a significant desertification effect in this area. It may be interesting to compare these areas with maps showing prior land uses, indicating which land uses contribute to large rates of desertification. Given this data on its own it is virtually impossible to separate extremely small-scale arid agriculture from other similar natural or urban environments. A further complication with this analysis is that urban areas generally have spectral characteristics similar to those of sandy soil, and therefore some image interpretation is needed to separate the various effects being observed. Additionally, if soil is varied, dark, or reddish in color, the use of a single sand endmember will not accurate depict increases in bare soil cover. In order to avoid these issues, a study region which was sparsely populated with predominately light sandy soil was used. |
|
For more information please contact Laurel_Firestone@yahoo.com |