Keywords

1 Introduction

Growing concerns over the recent urban and rural developmental activities, changing landscape scenario and the loss of biodiversity has made it an imperative for land managers to seek for better ways of managing the landscape and to be able to analyze positive and negative factors that influence the land use and land cover change pattern and its change dynamics. Land use land cover change is a continuous process which is altered by several ways with respect to time and space. The natural and socio–economic factors and their utilization by man in time have greatly affected historical land use and land cover pattern. Therefore, the accurate information pertinent to land use land cover is essential for the selection, planning and implementation of land use programmes which can be used to meet the increasing demands for basic human needs and their welfare (Zubair 2006). Rapid population increment and industrial sprawl has made land a scarce resource, mainly in urban and urbanizing areas which are undergoing to large–scale land transformation changes which alters the natural ecosystem in different way at temporal and spatial scale (Morley and Karr 2002). Change in land use land cover also aggravates the soil erosion, creates strong environmental impacts, effects on agricultural production, infrastructure and water quality (Lal 1998; Pimentel et al. 1995). The productive agricultural land and essential forest, neither of which can resist nor deflect the overwhelming land use and land cover but accelerate its change and alteration on the basis of necessity and sustainable development. This growth is an indicator of socio-economic development and generally has a negative enviornmental impact on the region. Agricultural practices, forest cutting, urban and industrial expansion, road development, military training, alteration of water way and other human activities have significant impact on land cover (Dale et al. 1998). Today one of the greatest challenge in front of human lies in the question of ‘how to minimize the negative impacts of land use land cover change by applying both appropriate and cost-effective technology respectively?’. The landscapes metrics is a special feature that provide the ability of quantifying land use and land cover pattern distribution. There are variety of landscape metrics to allow quantitative assessment of a landscape and its level of fragmentation (McGarigal and Marks 1995).

The analysis of landscape level and class level metrics Szabó et al. 2012) has provided a strong conceptual and theoretical basis for understanding landscape structure, function and change. The landscape metrics mainly focus on the three characteristics of landscape (Forman and Godron 1986; Turner 1989). (1) Structure: the spatial relationships among the distinctive ecosystems or elements present—more specifically, the distribution of energy, materials, and species in relation to the sizes, shapes, numbers, kinds, and configurations of the ecosystems. (2) Function: the interactions among the spatial elements, that is, the flow of energy, materials, and species among the component of ecosystems. (3) Change: the alteration in the structure and function of the ‘Ecological Mosaic’ over time. The landscape level metrics is a special feature of land use that provide significance ability to quantify the landscape patterns and the interactions among patch density, number of patches, total area, and large patch index within a landscape mosaic and also represent the patterns and interaction change over time. The class level metrics comprising with different metrics viz. core area, number of patches, patch density, percentage of landscape and largest patch index, some of these metrics quantify the landscape composition and while others quantify landscape configuration. It is also known that composition and configuration can affect ecological processes independently and interactively. So it is necessary to understand each metrics of landscape.

2 Study Area

This study area Allahabad district is situated at (24°47′N and 25°47′N to 81°19′E and 82°21′E) and it covers an area 5,246 km2. It lies in the northern-central part of the densely populated eastern region of state Uttar Pradesh, in between the Gangetic plain and adjoining Vindhyan plateau of India. River Ganga and Yamuna flowing through the district (Fig. 1 location map of the study area). Geologically the district presents a greater complexity, the whole Trans-Ganga tracts, the greater portion of ‘doab’ are composed of gangetic alluvium. The Trans Yamuna region is mainly quartzitic in nature and consisting denudational hills. Administratively the district comprised of eight tehils and twenty blocks. The mineral products that are commonly found in the district are glass sand, building stone, kankar, brick earth and reh. Building stone is extracted either by blasting or by splitting the chief quarries. Bricks and pottery, earth-material is available in the alluvial tract of the district and is locally used for the manufacture of bricks and earthenware.

Fig. 1
figure 1

Study area location map

Full size image

3 Materials and Methods

Total fourteen Survey of India (SOI) topographical sheet at 1:50,000 scales were used 63—G/10, G/11, G/12, G/14, G/15, G/16, H/13, K/2, K/3, K/4, L/1, K/6, K/7, K/8, and L/5 and satellite images of Landsat TM, 1990, ETM+ 2000 (http://www.usgs.gov/pubprod/aerial.html#satellite) and 2010 were used.

3.1 Classification of Satellite Data

According to Lu and Weng (2007), the major steps of image classification may include a suitable classification system, selection of suitable classification approaches, post-classification processing and accuracy assessment. These images were first geometrically and radiometrically corrected ERDAS IMAGINE 9.1 tool (ERDAS Field Guide 1999). The automatic atmospheric calibration was performed on each image separately. The Anderson et al. (1976) classification method was used as classification scheme. Afterwards, the Maximum Likelihood Classification tool is considered for image classification of year 2010 as it is taking account of both the variances and covariances of the class signatures and assigning each cell to one of the classes represented in the signature file. The algorithm used by the Maximum Likelihood Classification tool is based on Bayes’ theorem and the equation used in MLC classification as shown in the Eq. 4. The image of year 1990 and 2000 was classified using Isodata clustering approach (unsupervised method). The random sampling method was used for the accuracy assessment. The Google image, ancillary data and field data was employed for the accuracy assessment.

In the study following Eqs. 1, 2, and 3 were used e.g.:

$$\% \;Cover\;Change = \frac{{Area_{i\;year\;x} - Area_{i\;year(x + 1)} }}{{\sum\nolimits_{i = 1}^{n} {Area_{i\;year\;x} } }} \times 100$$
(1)
$$\% \;Cover\;Coverage = \frac{{Area_{i\;year\;x} }}{{\sum\nolimits_{i = 1}^{n} {Area_{i\;year\;x} } }} \times 100$$
(2)
$$\;Change\;area = Area_{i\;year\;(x + 1)} - Area_{i\;year\;x}$$
(3)

where \(Area_{i\;year\;x}\) area of is cover i at the first date; \(Area_{i\;year\;(x + 1)}\) is area of cover i at the second date; \(\sum\nolimits_{i = 1}^{n} {Area_{i\;year\;x} }\) is total cover area at the first date; \(t_{years}\) is period in years between the first and second scene acquisition dates.

$$D = \ln (a_{c} ) - \left[ {0.5\ln \left( {\left| {\text{cov}_{c} } \right|} \right)} \right] - \left[ {0.5({\mathbf{X}} - {\mathbf{M}}_{c} )\text{T} (\text{cov}_{c} - 1)({\mathbf{X}} - {\mathbf{M}}_{c} )} \right]$$
(4)

where, D is weighted distance; c is a particular class; X is the measurement vector of the particular pixel; Mc is the mean vector of the sample of class; ac is percent probability that any particular pixel is a member of class c; (Defaults to 1.0); Covc is the covariance matrix of the pixels in the sample of class c; |Covc| is determinant of Covc; Covc − 1 is inverse of Covc; ln is natural logarithm function; T = transposition function.

3.2 Fragmentation Analysis

A spatial pattern analysis program i.e., FRAGSTATS3.3 offers a comprehensive choice of landscape metrics and have been used to quantify landscape structure. It is implemented by decision maker and ecologists to analyze landscape fragmentation and to describe the characteristics and components of those landscapes. These statistics facilitates the comparison of landscapes and the evaluation of processes. The advantage of FRAGSTATS is that the calculations are implemented in a fully integrated fashion in a GIS and consequently easy to apply to digital map. The three indices have been taken into account for the ecological metrics analysis viz. patch density (PD), mean patch size (MPS) and percentage of landscape (PLAND). The number of patche per 100 ha, mean patch size (MPS) means the area occupied by a particular patch type divided by the number of patches of that type (McGarigal and Marks 1995) and percentage of landscape (PLAND) (McGarigal and Marks 1995) measure of landscape composition. The representative equations for all the three indices are shown by following Eqs. 5, 6, and 7.

$$PD = \frac{{n_{i} }}{A}(10{\text{,}}000)(100)$$
(5)

where \(n_{i}\) is number of patches in the landscape of patch type (class) i; A is total landscape area (m2).

$$MPS = \frac{A}{N}\left( {\frac{1}{10{\text{,}}000}} \right)$$
(6)

where N is total number of patches in the landscapes, excluding any background patches

$$PLAND = \frac{{\sum\nolimits_{j = 1}^{n} {a_{ij} } }}{A}(100) = P_{i}$$
(7)

Pi is proportion of the landscape occupied by patch type (class) i; aij is area (m2) of patch ij.

4 Result and Discussion

4.1 Land Use and Land Cover Distribution

The study area has been classified into different land use classes shown in Fig. 2. The overall accuracy for the classified images was arrived 80.39, 85.36 and 88.10 % respectively of years 1990, 2000, and 2010. According to Anderson et al. (1976), the accuracy of classified images should >84 % for better results. On the basis of analysis the result showed that the builtup area was 555.89 km2 in 1990 and it increases 647.80 km2 in 2000 and it further increases to 744.16 km2 in 2010 respectively. This may be attributed to an increase in urban population. As the population of the district shows the increasing trend of population in the region. The area of agriculture land showed nominal increment from 3,638.40 km2 in 1990 to 3,739.22 km2 in 2000 and 3,793.25 km2 in 2010, which means that the area remained same for agriculture. Cultivable area decreased from 246.19 km2 in 1990 to 193.28 km2 in 2000 and it further decreased to 116.97 km2 in 2010 (Table 1). This decrease in cultivable area may be due to either urban pressure or nominal conversion into agriculture land. The difference in land use classes are given in (Table 2) which shows change in area which indicates that in the year 2000/1990 the forest is −86.89 km2 while in the 2010/2000 its area is increased by 62.94 km2, however in overall analysis it showed 2010/1990 it reduced −23.95 km2 respectively indicates that the fluctuations in forest cover area could be due to social and economic influence. Built up area change 2000/1990 are 91.91, in 2010/2000 are 96.36 and in 2010/1990 increase in 188.3 km2 indicates that the built up area continuously increased over the study period.

Fig. 2
figure 2

Unsupervised and supervised classified satellite images of year 1990, 2000, and 2010

Full size image
Table 1 Area of different land use classes in km2 for the year 1990–2000–2010
Full size table
Table 2 Difference in area of different land use classes during 1990–2000–2010 (positive values shows decrease while negative values shows increase)
Full size table

The change area of cultivable land in 2000/1990 was −52.91, in 2010/2000 was −76.31 and in 2010/1990 was −129.2 km2. It indicates a slight decline of cultivable land due to urbanization. It indicates that this change may be due to agricultural and built-up area expansion. The agriculture area in 2000/1990 is 100.8 km2, while reduced in 2010/2000 is 54.3 km2 and however it further showed increment in 2010/1990 is 154.9 km2 respectively. Water body area in 2000/1990 is −3.9, while in 2010/2000 its area is reduced by −26.18; however in 2010/1990 again it shows reduction by −30.08 km2 respectively, which indicates that the change in water body area could be due loss of surface water body.

4.2 Landscape Level Metrics

Landscape metrics were characterized for the landscape fragmentation patterns which reveals the configuration and composition pattern of the landscape element such as in the form of class, patch and landscape metrics. The spatial characteristics of patch, class of patches or entire landscape also quantifying and exploring with the help of ecological process by landscape metrics (Narumalani et al. 2004). The landscape level metrics analysis analyses the following parameters i.e. total area (TA), number of patches (NP), patch density (PD), and Area mean. PD metrics is an important measure to show the health of an ecosystem. If the NP increases while the area under the class decreases, it represents fragmentation or dissection. Anthropogenic and natural activities affect the spatial structure of landscape such as urbanization and industrialization due to anthropogenic activities while the natural activities like flood and drought etc. The nature of these change experienced by each land use land cover varies.

The landscape can undergo in different type of transformation, the result after the change process analysis are given in Table 3 shows that the total area 499,182.5 ha in 1990, 498,818.6 ha in 2000 and 550,046.2 ha in 2010. The number of patches per unit area has increased from 3,301 ha in 1990 to 42,394 ha in 2000 and further it decreases by 4,646 ha in 2010 respectively in twenty block of study area. The increase in number of patches in the first decade shows that the landscape is more fragmented while in the second decade it shows that the landscape gets less fragmentation. The maximum number of patches (582 in year 2010) found in Koraon block. Kaurihar and Soraon have number of patches 240, 134 in 2000 means these areas are fragmented during the year 2000. While low number of patches i.e. 54 was present in Dhanupur means this area has less developmental activities. PD increases from 12.755 in 1990 to 386.142 in 2000 and after one decade it rapidly decreases to 19.821 in 2010. These values indicate that the landscape fragmentation was at higher pace during the year from 1990 to 2000 while it was less fragmented during the year in 2000–2010. These change in values of mean area indicates that the forest is less fragmented in 2000 than in 1990 and again it is more fragmented in 2010. Two decades of natural conditions or human activities had less impact on forest landscape and thus fragmentation was less in 2000. While in 2010, human pressure or natural conditions may played a major role in the decrease of forest mean area.

Table 3 Landscape level metric analyses during 1990–2000–2010
Full size table

4.3 Class Level Metrics

The class level metrics has a spatial feature to represent each land use land cover classes. The importance of class level metric analysis is to assess the transformation types which affect the spatial pattern of the landscape. In the present study we calculated the core area (CA), number of patches (NP), patch density (PD), and percentage of landscape (PLAND), and Area mean of land use classes. The class level metric analysis shows that CA of agriculture land (Table 4) from 345,595.3 ha in 1990 to 363,219.9 ha in 2000 which further decreases 334,817.4 ha in 2010. Out of twenty blocks the maximum core area of agricultural land was found in Koraon block 43, 684.5 in 1990, 43,683.94 in 2000 and 42,928.23 in 2010.ha, while minimum core area of agricultural land was reported in Kaundhiyera (1990) and Soraon (2010) 7,759.75 and 8,065.66 ha respectively. The overall study reveals that the year 2000 comprises the maximum core area in two decades. The number of patches of agricultural land decreases from 550 in 1990 to 508 in 2000 and while it increases 799 in 2010. Out of twenty blocks NP was more in Chaka (2010), Kaundhiyera (2010) block 89 and 84 respectively. The low NP of agricultural land was found in Dhanupur (2000), Pratappur (2000) block 7 and 9 respectively, the low NP in the year 2000 shows less fragmentization of agricultural land compared to 2010. The PD was 2.18 in year 1990 of agricultral land which decrease to 1.78 in year 2000, and again increases to 3.18 in 2010. Maximum PD was found in Soraon (2010) 0.39 and Kaundhiyera (2010) 0.38 while minimum in Dhanupur and Pratappur (2000) 0.04 and 0.04 respectively. PLAND in two decade’s study was found maximum in Dhanupur (1990) 93.85 and Handia (2000) 87.48, while it was minimum in Chaka (2010) 4.88 and Kaundhiyara (1990) 35.36 out of twenty block. The Area mean of agriculture land firstly increased from 14,588.2 ha in 1990 to 17,742.5 ha in 2000 and then decreased to 11,036 ha in 2010, the maximum area occupied by Dhanupur (2000) and Pratappur (2000) was 2,212 and 2,099 ha respectively. The minimum in Chaka (2010) and Kaundhiyara (2010) was 24.58 and 101.7 ha respectively.

Table 4 Class level metric analysis for core area, number of patch, patch density, percentage of land, and area mean (ha) during 1990–2000–2010 of agricultural land
Full size table

The CA of built up area (Table 5) decreases from 25,483.8 ha in 1990 to 24,783.5 ha in 2000 and it again increases from 24,783.5 to 38,045.7 ha in 2010. Out of the twenty block maximum CA was found in Chaka (2000) 8,485.91 ha and Bahadur Pur (2010) 3490.648 ha while minimum CA was found in Bahadurpur (1990) was 270.97 and 274.7842 ha in Kaurihar (1990). The two decade’s study, showed that the year 2010 has more CA 38,045.7 ha compared to 25,483.8 ha in year 1990 of builtup area. The NP increased from 741 in 1990 to 775 in 2000; it again increased to 1863 in 2010. Shankargarh (2010) and Jasra (2010) bock have more NP 240 and 212 respectively. The result indicates that the year 2010 is fragmentized as compared to previous decades. The PD has decreased from 3.4 ha in 1990 to 3.21 ha and in 2000 it further increased to 7.26 ha in 2010. PD in Holagarh (2010) 0.78 and Mau (2010) 0.56 are found at maximum but a minimum in Bahadurpur (1990) was 0.07, Urwa (2000) and Shankargarh (1990) were 0.07. The lowest PD was obsevered in Meja 0.5 in year 2000.

Table 5 Class level metric analysis for core area, number of patch, patch density, percentage of landscape and area mean (ha) during 1990–2000–2010 of built up area
Full size table

The PLAND also decreased from 95.07 ha in 1990 to 80.89 ha in 2000, while again increased 146.6 ha in 2010. The maximum value of PLAND was found in Soraon (2010) 15.63 ha and Chaka (2000) 18.99 ha while minimum is 1.37 ha in Bahadurpur (1990) and 1.048  ha in Meja (2000). The Area mean of built-up area increases from 689.5 ha in 1990 to 799.9 ha in 2000 and again decreases to 611.32 in 2010. Out of twenty blocks the maximum Area mean contained by Chaka 292.6 ha in 2000 and minimum in Soraon (2000) and Jasra (2010) with 6.64 and 7.55 ha respectively.

The CA for cultivable area (Table 6) decreases from 18,699.62 ha in 1990 to 14,696.69 ha in 2000 and remains to 14,696.69 ha in 2010. The maximum CA is investigated in Jasra (2000) 7,983.1 ha and Kaundhiyara (2000) 9,530.8 ha, while minimum CA was found in Dhanupur (1990) 0.0812 ha. The NP of cultivable area was found 217 in 1990 and it increases to 422 in 2000 but further it shows reduction to 311 in 2010. Out of twenty blocks maximum NP was found in Koraon (2000) 65 and minimum 1 in Bahadurpur, Dhanupur and Kaurihar (1990) respectively. The PD also decreases in two decades study period from 0.29 in 1990 to 0.27 in 2000 and 0.22 in 2010, while the block Holagarh (2000) and Kaundhiyara (1990) has maximum 0.17 PD and Dhanupur (1990), Kaurihar (1990) and Baheriya (2010) contained minimum PD 0.01. But in two decades study period the year 1990 contained more 0.29 PD. The PLAND also decreases from 105.73 in 1990 to 41.606 in 2000 and further increases 350.22 in 2010. But the maximum value was found in Urwa (1990) 81.69  and minimum in Mau (1990) 0.001. In the two decades study reveal that PLAND was high in the year 1990 in cultivable land. The Area mean of cultivable land found in increasing order from 1,567.69 ha in 1990 to 2,037.17 ha in 2000 and it further increases from 2,037.17 to 3,312.62 ha in 2010. The maximum Area mean are 804.7 ha in Chaka (2010) and Jasra (2010) 506 ha. But the minimum Area mean are in Saidabad (1990) and Phulpur (2010) 0.045 and 0.1 ha respectively.

Table 6 Class level metric analysis for core area, number of patch, patch density, percentage of landscape and area mean (ha) during 1990–2000–2010 of cultivable area
Full size table

The CA of forest area (Table 7) shows increase from 65,311.38 ha in 1990 to 78,226.87 ha in 2000 but rapid decrease of 37,260.71 ha in 2010. Out of the twenty blocks the maximum CA are Koraon (1990) 11,982 ha and Karchhana (1990) 9,146 ha but the minimum are Mau, Phulpur and Chaka (2010) 0.223, 0.39 and 0.167 respectively. This result indicates that developmental activity and human pressure put a great impact in forest area reduction over the last two decades. The NP of forest area decreased from 250 in 1990 to 158 in 2000 and further it continues reduction to 141 in 2010. The maximum NP were found in Koraon 22 (1990) and Handia 17 (1990) and minimum in Dhanupur (2010) 1 and Saidabad 1 (1990) PD of forest area in 1990 was 1.512 and increases 4.53 in 2000, it then decreases 1.61 in 2010. The maximum PD was found in Chaka (2010) 0.98 and Urwa (2000) 0.987. The minimum PD in Saidabad 0.005 (1990) and 0.007 in Meja (2010) was obtained. The value of PLAND drastically decreases from 1,333.72 ha in 1990 to 43.10 ha in 2000 and further increases 95.39 ha in 2010. Out of twenty block the maximum PLAND 18.15 ha in Manda (1990), 18.12 ha in Koraon (1990) and 10.26 ha in Karchhana (2010) but the minimum PLAND was found at 0.016 ha in Holagarh (2000) and 0.054 ha Bahadurpur (2000). The Area mean accounts in 1990 is 8,048.71 ha which decreases to 3,131.77 ha in 2000 and goes on to increase in 2010. The maximum Area mean reported out of twenty in Jasra (1990) and Holagarh (1990) 891.5 and 788.7 ha respectively.

Table 7 Class level metric analysis for core area, number of patch, patch density, percentage of landscape and area mean (ha) during 1990–2000–2010 of forest area
Full size table

The CA of water body (Table 8) decrease from 34,568.5 ha in 1990 to 22,580.5 ha in 2000 and gradually increases to 37,287.5 in 2010. The CA was maximum in Chaka block (2010) 11106  ha and Koraon (1990) 3,730 ha. After two decades study revealed that the CA was found maximum in year 2010 compare to year 1990. The NP of water body was 1,071 in 1990 and it decreases 982 in 2000 and further increases 1,234 in 2010. The maximum NP was found 176 in Koraon block 1990 and minimum 7 in Holagarh in 2000. The PD in 1990 was 3.91 in 1990 and decreases 3.55 in 2000 and further increases to 4.67 in 2010. The PLAND decreases from 40.65 in 1990 to 22.66 in 2000 and increases to 31.47 in 2010. In two decades the data showed that PLAND was high in 1990 and low in 2000. But out of twenty block the value of PLAND was found higher in Meja (1990) 13.12 ha and Urwa (1990) 10.52 ha but minimum was 1.72 in Holagarh (2000), Dhanupur 0.95 in 2000. The Area mean of water body also increases from 1,345.44 ha in 1990 to 593.5 ha in 2000 and further increases 613.38 ha in 2010. Out of twenty block the maximum Area mean occupied by Chaka (2010) and Bahadurpur (2000) 137.1 and 87.23 ha respectively.

Table 8 Class level metric analysis for core area, number of patch, patch density, percentage of landscape and area mean (ha) during 1990–2000–2010 of water body
Full size table

4.4 Class Level Metric Analysis

The results from class level metric analysis are discussed in this section.

Percentage of landscape (PLAND) is a unique technique which revealed landscape composition, the different landscape have different land use pattern with unique attributes like residential land use has high patch density and agriculture land use has high mean patch size (Weng 2007). The PLAND of built-up area of selected blocks are positively related to the developmental activities like commercial, transportation, industrial development and human intervention. In this study topographical factors have great impact on the change of PLAND of cultivable land. The PLAND of forest and agriculture land are found to have decreased. The PLAND of water body and built-up area also change in two decades showing in Fig. 3.

Fig. 3
figure 3

Change of class level percentage of landscape (PLAND) of different land use class of a built-up, b cultivable, c agriculture, d forest and, e water body

Full size image

The area mean of class level metrics denote the average patch size of all patch belonged to the same land use type in landscape (Fig. 4) The area mean of urban land use type (built-up) is high compared to rural land use type (agriculture, forest and water body). The change in area mean of agriculture gradually decreases in two decades while the built up and cultivable land is positively correlated with the degree of human intervention and commercial practices. The PD is expressed by the number of patch per 100 ha. PD of any particular type of landscape denotes the current condition of all land use type metrics showing Fig. 5. The change of PD of different land use type like agriculture, built up, cultivable, forest and water body have different values compared to each year differently which revealed that the fragmentation of landscape has occurred due to human intervention and developmental activity. PD of overall twenty blocks among the five landscapes like forest area has more PD than built up, cultivable, agriculture and water body. The unusual high PD of forest area at Baheriya and Soraon block are due to road, household and commercial extension. The PD of forest and agriculture area has increased whilst the PD of built-up, cultivable and water body decreases in previous two decades gradually.

Fig. 4
figure 4

Change of class level: area mean (Area Mean) of different land use class of a built up, b cultivable, c agriculture, d forest and, e water body

Full size image
Fig. 5
figure 5

Change of class level: patch density (PD) of different land use class of a built up, b cultivable, c agriculture, d forest and, e water body

Full size image

The NP provides clear information about the class level metrics high value indicates more fragmentation in particular area. The NP of agriculture and forest land is more as compared to cultivable, built-up and water body. That reaveled the change in NP  which is highly correlated with the human intervention showing in Fig. 6. The core area of the twenty blocks has greatly influenced in the past  two decades which reveal the continuous change of class level metrics. Among the five class showing in Fig. 7, the agriculture and cultivable land have more CA compare to built-up, forest and water body in current time.

Fig. 6
figure 6

Change of class level: number of patch of different land use class of a built up, b cultivable land, c agriculture, d forest and, e water body

Full size image
Fig. 7
figure 7

Change of class level: core area of different land use class of a built up, b cultivable, c agriculture, d forest and, e water body

Full size image

The patterns of change of class level metrics are more complicated. Generally, the results of built-up area and agriculture area are similar with metric outputs. This indicates a positive relationship to the degree of human intervention and developmental activity due to population pressure. The cultivable, forest and water body has also affected by human intervention.

4.5 Land Fragmented Class Analysis

Figure 8 shows the changes of landscape pattern both spatially and temporally through two decades. In the graph vertical axis represents the value of class and the horizontal axis represents the name of twenty blocks. TA mainly comprises of the overall area which is going through different processes of land use change. The TA value of land fragmented class when compared to previous two decades reduced and converted into many land use class with respect to the degree of human interference, as shown in Fig. 8a. The area mean shown in Fig. 8b presents gradual decrease within the last two decades indicating a fragmented area mean. Temporally, area mean in most of the blocks of land fragmented class is previously seen as decreasing and then it increases, which reveals the reciprocally changed area mean that highly correlates to the degree of human interference and developmental activity

Fig. 8
figure 8

a Total area. b Area mean. c Patch density and, d Number of patches in different land fragmented class

Full size image

PD at high value are indicating a fragmented landscape (Fig. 8c), which is directly or indirectly effected by the topographical or anthropogenic activity. Theoretically the observed changes of PD indicating positive relationship to the degree of commercial transportation (road construction) and residential expansions. Temporally, PD has increased through time in most of the blocks, indicating an increasing fragmentation of land fragmented class. The high value of number of patch in a particular landscape denotes the most fragmentation. Thus the NP has significant information of current land use practice. Overall study of twenty blocks reveals that increase of NP value frequently in the current year which represents explicit information of fragmentation of landscape due to human intervention and developmental activity.

4.6 Estimating Effects on Water Quality

Rapidly growing countries, and particularly those emerging from rural to urban, frequently lack the resources to adequately monitor and forecast the impacts of land development on biologic, hydrologic, and social resources. Urbanization has significantly changed natural landscapes everywhere (Forney et al. 2001). Urban growth and fragmentation caused by urban sprawl have been extensively studied (Herold et al. 2002; Gonzalez-Abraham et al. 2007). Landscape structure is one of the most important factors influencing nutrient and organic matter runoff in watersheds (Turner et al. 2003; Wickham et al. 2003; Uuemaa et al. 2007). Therefore there is increasing demand for indicators and methods that make it possible to evaluate the landscape factors influencing water quality in freshwater management (Griffith 2002). Several studies have attempted to determine the relationship between land use/land cover structure and water quality but most studies have largely relied on compositional landscape metrics (Kearns et al. 2005). It is, however, clearly important to understand not only the total area of sources and sinks in the landscape, but also their spatial arrangement relative to flowpaths. The importance of the spatial arrangement of land cover within watersheds on water quality has been studied by Jones et al. (2001), King et al. (2005), Xiao and Ji (2007), Uuemaa et al. (2005, 2007).

The groundwater after analysis showed that the areas which are more urbanized and fragmented have elevated concentration of different ions in the water. Three sites, namely Koran, Bahariya and Karchhana, were found to have elevated fluoride in analyzed groundwater samples. Similarly, Urwa, Holagarh and Soraon had positive value of fluoride so water is not suitable for drinking. The fluoride in pre-monsoon found that Koraon, Chaka (Naini) and Bahariya had values of fluoride of 1.096–1.291 mg/l and Jasra, and Bahadurpur had moderate values of fluoride (Singh et al. 2002). The domestic pollution is the major source of pollution in Yamuna River. About 85 % of the total pollution in the river is caused by the domestic sources. About 3 km upstream from confluence with Ganga River at Grand Trunk Road Bridge called Naini Bridge near Gau Ghat (MoEF 2006).This location depicts the Yamuna River water quality before its confluence with river Ganga. The domestic pollution is mainly caused by the urban centers. The organic pollution and microbial contamination reflect increasing trend up to Allahabad. The yearly average from 1999–2005 showed that the Dissolved oxygen was highest during year 1999 and 2000 (8 mg/l) afterwards it showed decreasing trend (6.5 mg/l). During the year 2004 and 2005 Biochemical oxygen demands was higher as compared to previous years. Chemical oxygen demand at Allahabad it ranged between 5 and 18 mg/l. These results showed that the Yamuna River quality is deteriorating due to organic pollution and this may also be attributed to landscape fragmentation.

According to Singh et al. 2002, heavy metal analysis in the <20-μm-fraction of stream sediments appears to be an adequate method for the environmental assessment of urbanization activities on alluvial rivers. In their study they sampled during the pre-monsoon season at Delhi, Agra, Kanpur, Allahabad and Varanasi in 1993 and at Lucknow in 1995. These wide ranges are attributed to differential behavior of heavy metals rich urban effluents draining into rivers of the Ganga Plain. Stream sediments from Allahabad are classified into Sediment Pollution Index (SPI) natural sediments. To understand the impact of urbanization activities on stream sediment quality, the Pollution Load Index (PLI) given by Tomlison et al. (1980) was calculated for each urban centre. Allahabad 1.13 this PLI values show direct linear relationship with urban centre population.

5 Conclusion

The results of our study are partially validated. This study presents the results of a set of landscape metrics derived from remotely sensed data aiming to characterize the historical trends of landscape changes in the Allahabad district in the period 1990–2010. However, the identified trends in landscape changes in the region have potential policy implications in the region. The present study clearly reveals the impact and influence of the action taken during prolonged phase of land use/cover monitoring followed by a relatively short phase of passive management of twenty block of study area of Allahabad region during past two decades. The study is taken up to characterize the landscape pattern using land use/cover, landscape metrics and class level ecological metric analysis type that affect pattern of landscape changes. The landscape metric analysis indicates that from the year 1990 to 2000 periods the fragmentation of landscape was slightly low because the natural and climate condition are probably good, while from the year 2000 to 2010 it deteriorates which indicates that it could be due to human induced disturbance and development to gain more luxurious life, this still needs to be validated. The study on landscape and class level metrics helps for land management officer, forest management officer and policy makers in assessing and quantifying the extent to which the land use/cover is affecting the landscape and thus provide a complete perspective of change. Since most of changes are occurred in this area are due to human intervention and thus study over the fragmentation pattern will help to understand the spatial landscape transform processes which are also important for the ecological system and services. Landscape metric and landscape transformation analysis showed that over the time spatial configuration and composition of the landscape has changed a lot which leads to the environmental degradation. This study demonstrates the probable use of remote sensing, GIS and FRAGSTAT in assessing spatial structure and change in landscape.