eChapters

Introduction

Over the past few decades, soil resources have witnessed many changes driven by demographic pressure and climate change. It is necessary to have up to date soil information to monitor the state of soils health with time. Soil resource inventory provides information about the potentialities and limitation of soils for its proper utilization. Soil survey offers an accurate and scientific inventory of different soils, their properties and extent. An in-depth knowledge of the type of soils and their spatial distribution is a prerequisite for developing rational land use planning. Remote sensing is the science of obtaining information about the earth’s surface without actually being in contact with it. On the other hand, proximal sensing refers to remotely sensed measurements that are taken at the field or laboratory level. Both these processes involve an interaction between incident radiation and the targets of interest. Fundamentally, remote sensing data products are a representation of the functional response of objects to energy sources. These are typically spectra, which are acquired as a discrete series of reflectance measurements taken at different wavelength intervals, or bands. In the shorter wavelengths of the electromagnetic spectrum (visible part), features can be detected by virtue of reflected solar energy, while in the longer wavelengths (microwave, thermal parts), sensing of emitted energy works. Conventional soil sampling, and subsequent laboratory analysis are generally time consuming and costly. In this context, remote sensing plays a crucial role for studying soil resources. The term remote sensing is generally used for airborne and space borne acquisitions, whereas proximal sensing refers to ground-based laboratory and field measurements. Advances in the quantitative disciplines like remote sensing and proximal sensing have laid the foundations for a spatial exploration of soil-system dynamics within a landscape context (Pennock and Veldkamp 2006).

Introduction

Soil is one of the vital resource, and is the basis of the existence of mankind. Management of this resource including its conservation and utilization is of crucial importance. It is a base on which all life depends. In the recent past, with burgeoning populations and the national goals of seeking self-sufficiency in food and fibre production, this resource base is slowly being stripped. While natural systems often adapt to stress in a remarkable fashion, some relationships - once destroyed - can never be restored. Soil resources can be assessed and monitored through its spectral signatures gathered through remote sensing.

The term ‘remote sensing’ is commonly restricted to methods that employ electromagnetic energy (such as lights, heat, micro wave) as means of detecting and measuring target characteristics. Sensors acquire data as various earth surface features reflect or e mit electromagnetic energy (EM energy). Remote sensing is the technology that is now the principal method (tool) by which the earth’s surface and atmosphere (as targets or objects of surveillance) are being observed, measured, and interpreted from vantage points. Remote sensing of the earth traditionally has used reflected energy in the visible and infrared, and emitted energy in the thermal infrared and microwave regions to gather radiation that can be analyzed numerically or used to generate images whose tonal variations represent different intensities of photons associated with a range of wavelengths that are received at the sensor.

Introduction

Accuracy and uncertainty analysis is an important step in preparation of surface map through geostatistical techniques or digital soil mapping approach. It is because that we rely on a map based on how much accurate it is and the degree of uncertainty associated with the estimates while the map was prepared. Any stakeholder for a soil map especially policy makers is interested in accuracy and uncertainty of the map before taking suitable land management decisions involving soil map as an input. Again the level of accuracy and uncertainty depends on the properties for which the map has been developed. If the property of interest is sensitive to environment or human consumption, then the accuracy level should be very high while the uncertainty level should be very low e.g. heavy metal concentration in soil, quality parameters of water etc.

Introduction

GIS (Geographical Information System) based mapping takes cogence of three major geometrical dimensions, point, line and polygons and in GIS these three features correlate their relationship through a concept called Topology. Among the various applications where covariate processing requires specific attention, the landuse/ land cover and Digital Elevation Models/surfaces or terrains are the major area of interest. Use and processing of data for these covariates carries importance as accuracy of both these aspects would influence the derivatives extracted out of them further. For example, if landuse is improperly defined, the site for which planning is proposed may not sustain, similarly, point data/elevation data in case of DEMs would decide the exact topography of any landforms for their further utilization. For their use in GIS, while DEMs depend upon point data, the landuse/ land cover will be taken care by polygons. For the purpose of the study, we are going to discuss about these two covariates in the following paragraphs.

Generally, Covariate (covariable) (koh-vair-iãt) in statistics is defined as a continuous variable that is not part of the main experimental manipulation but has an effect on the dependent variable. In other words, as per ANCOVA (Analysis of covariance) definition: a covariate is a control variable.

Introduction

Quantum Geographical Information System (QGIS) is an open source GIS environment. The software package is available freely in web https://www.qgis.org/en/site/forusers/download.html) and can be installed in a variety of computer platform e.g. Windows, Linux, Mac OS X, Android etc. Latest information on QGIS is available inhttp://www.qgis.org/en/site/. Basically, the QGIS GUI is divided into five areas: (i) Menu Bar, (ii) Tool Bar (iii) Map Legend, (iv) Map View and (v) Status Bar. QGIS allows users to define a global and project-wise coordinate reference system (CRS) for layers without a pre-defined CRS. It also allows the user to define custom coordinate reference systems and supports on-the-fly (OTF) projection of vector and raster layers. All of these features allow the user to display layers with different CRSs and have them overlay properly. QGIS has support for approximately 2,700 known CRSs. Definitions for each CRS are stored in a SQLite database that is installed with QGIS. The CRSs available in QGIS are based on those defined by the European Petroleum Search Group (EPSG) and the Institute Geographique National de France (IGNF) and are largely abstracted from the spatial reference tables used in Geospatial Data Abstraction Library (GDAL). For Post GIS layers, QGIS use the spatial reference identifier that was specified when the layer was created. For data supported by OGR, QGIS relies on the presence of a recognized means of specifying the CRS. In the following sections, few basic GIS analysis using example spatial data from arid western India is discussed. However, readers are advised to go through different tools and functions available in QGIS after installation of the system in personal computers.

Introduction

Spatial datasets always exhibit two common features (i) the occurrence of a few very large concentrations (hot-spots) and (ii) the presence of data below the detection limit (black spots). Extreme values can strongly affect the characterization of spatial patterns, and subsequently the prediction. Several approaches exist to handle strongly positively skewed histograms (Saito and Goovaerts 2000). One common approach is to first transform the data (e.g. normal, Box Cox or lognormal), perform the analysis in the transformed space, and back-transform the resulting estimates. Such transform, however, does not solve problems created by the presence of numerous black spot data, since either it yields a spike of similar transformed values or, in the case of the normal-score transform, it requires a necessarily subjective ordering of all equally-valued observations. Moreover, except for the normal score transform (Deutsch and Journel 1998), it does not guarantee the normality of the transformed histogram, which is required to compute confidence intervals for the estimates. Last, the back-transform of estimated moments is not straightforward and can introduce bias if not done properly (Saito and Goovaerts 2000); for example, lognormal kriging estimates cannot simply be exponentiated. Another way to attenuate the impact of extreme values is to use more robust statistics and estimators. The non-parametric approach of indicator kriging (IK) falls within that category (Journel 1983, Goovaerts 2001). The basic idea is to discretize the range of variation of the environmental attribute by a set of thresholds (e.g. deciles of sample histogram, detection limit, regulatory threshold) and to transform each observation into a vector of indicators of non-exceedence of each threshold. Kriging is then applied to the set of indicators and estimated values are assembled to form a conditional cumulative distribution function (ccdf). The mean or median of the probability distribution can be used as an estimate of the concentration of the material in question (Barabas et al. 2001, Cattle et al. 2002, Goovaerts et al. 2005).

Introduction

In geographical location, the observations taken in proximity tend to be more alike than observations made at points that are far apart-that is, the values tend to be correlated. An area of study that takes this sort of spatial consideration into account is called geostatistics. A primary motivation for sampling is to make meaningful estimates of values at nearby positions in space and time. When values are assumed independent, the best estimate for an unmeasured point will be the mean. However, for correlated values nearby values are estimated by interpolation. The simplest interpolation scheme is a nearest neighbor estimate for which the unknown values are based on the closest measured location. Inverse distance is another sort of interpolation technique where weights are estimated based on the inverse distance squared. Kriging is a geostatistical interpolation technique that considers both the distance and the degree of variation between known data points when estimating values in unknown areas.

It makes the best use of existing knowledge by taking account of the way that a property varies in space through the variogram model. In its original formulation a kriged estimate at a place was simply a linear sum or weighted average of the data in its neighbourhood. Weights are based on the distance between the measured points, the prediction locations, and the overall spatial arrangement among the measured points. Since then kriging has been elaborated to tackle increasingly complex problems in mining, petroleum engineering, pollution control and abatement, and public health. The term is now generic, embracing several distinct kinds of kriging, both linear and non-linear.

Introduction

Spatial variation structure of soil properties is a key input for digital soil mapping and generally determined through geostatistical techniques (Webster and Oliver 2007). In geostatistics, spatial variation is expressed by semivariogram ?(h) , which measures the average dissimilarity between data separated by a vector h. It is generally computed as half of the average squared difference between the components of data pairs:

Introduction

The first step of geostatistical analysis in R is the creation of dataset which is acceptable in R environment by using the data collected from field survey or laboratory experiments. Different type of data is collected during experiments e.g. sand content, silt content, organic carbon content, geographical location of data points, soil type, land use type etc. Here, we will discuss how these data are converted into a acceptable dataset format for R environment and how we manage these data. In R this task is performed in two steps. First, a data structure is created to hold the data and second is the entering or exporting the data into data structure.

Introduction

Spatial data is different from a non-spatial data in the sense that it always has a spatial reference system which is nothing but a pair of coordinate values and a system of reference for these coordinates. For example, the locations of benchmark soil series of India can be defined as pairs of longitude and latitude in decimal degree values with respect to the prime meridian at Greenwich and zero latitude at the equator. The World Geodetic System (WGS84) is a frequently used system of reference on the Earth. The different types of spatial data are: (i) point, which is a single point location, such as a GPS reading or a geocoded address (ii) line, which is defined as a set of ordered points, connected by straight line segments (iii) polygon, an area, marked by one or more enclosing lines, possibly containing holes and (iv) grid, a collection of points or rectangular cells, organised in a regular lattice. Among the above said spatial data, the first three are vector data models and represent entities as exactly as possible, while the fourth data model is a raster data model, representing continuous surfaces by using a regular tessellation. In this chapter, we discuss how a spatial data can be handled in R. There are several packages available in R in which spatial data analysis can be handled e.g. sp, rgdal, raster etc. All the contributed packages for spatial data analysis in R address two broad areas: moving spatial data into and out of R, and analysing spatial data in R. In the following section, the sp package for spatial data handling is briefly discussed.

Introduction

R is one of the most popular platforms for data analysis and visualization currently available. It is free, open-source software, with versions for Windows, Mac OS X, and Linux operating systems. R is a language and environment for statistical computing and graphics, similar to the S language originally developed at Bell Labs. It’s an open source solution to data analysis that’s supported by a large and active worldwide research community. But there are many popular statistical and graphing packages available e.g. Microsoft Excel, SAS, IBM SPSS, Stata, and Minitab etc., however R has many interesting features so why it is mostly followed by researchers and academicians. Few of the major features of R are given below:

Soil sampling is an integral and essential component of soil fertility evaluation and nutrient management research. The effectiveness of soil sampling is a prerequisite for soil testing to achieve its goals for efficient nutrient management and plant nutrition for improved yield and quality. The underlying basis for soil sampling is that a soil sample taken represents the “population” which may be a plot, field or a watershed. It further implies that nutrient status of the representative soil sample(s) determined in a laboratory would reflect nutrient status of a plot, field or watershed and is of interest for correcting nutrient disorders in the field or watershed.

The most important factor that influences the effectiveness of soil sampling is soil heterogeneity. However, in a relatively homogenous group of fields or plots, a small number of samples may be sufficient to represent the population to a more heterogenous group of fields that would require more number of samples to represent the soil population. Whenever, soil samples to be taken in field following points needs to be kept in mind.

The arid region in India is spread over in 38.7 million hectare land. Out of the total 31.7 m hectare lies in hot arid region and remaining 7 m hectare comes under cold arid region. The hot arid region occupies major part of northwestern India (28.7 m ha) and a small pocket (3.13 mha) in southern India. The north western arid region occurs between 22°30’ and 32°05’ N latitude and 68°05’ to 75°45’ E, covering western part of Rajasthan, north western Gujarat and south western parts of Haryana and Punjab. About 62% area of arid region falls in western Rajasthan and 20% in Gujarat. Haryana and Punjab together constitute 7% area of arid region. The state wise distribution of arid region is shown in Table 3.1. Further our discussion is based on the majority area of arid region, belonging to Rajasthan, Gujarat, Punjab and Haryana.

Introduction

A complex spatial soil properties pattern is directly controlled by soil internal factors and anthropogenic impacts.Hartemink et al. (2001) elaborate a strong decline in mineralogy, morphology and genesis of soil research than strong increase in pedometrics applications during 1967 to 2001. Hartemink and McBratney (2008) identified this change as “soil science renaissance”. Basically, the soil science deeply rooted in agriculture, geology, and chemistry but the paradigm has been shifted from classification and inventorization to quantification of soil patterns in spatial as well as temporal patterns and their impact on the ecosystem health and hydrological cycle. The environmental keen approach explains that soil is in the center of the ecosystem and highly interacting with the biotic and abiotic factors of the ecosystem and a complex pattern and process has been co-evolved with the time.

Introduction

Geostatistics is a branch of statistics that deals with the analysis and modelling of geo-referenced data. Its main aims are to quantify spatial variability and to create maps from point observations. Of course there is much more to geostatistics than just that and indeed some of the other uses of geostatistics are addressed in this chapter too, but the emphasis is on the assessment of spatial variability and geostatistical interpolation.

The first step of geostatistical interpolation is to quantify the spatial correlation structure of the variable of interest. This can be done by examining the observations and how these vary in space. Next spatial interpolation makes use of the quantified spatial correlation to derive optimal predictions at unobserved locations and create a map. The interpolation error is quantified as well, which helps to design optimal spatial sampling schemes that balance data collection costs and map accuracy. All this will be explained in this chapter, but in order to do so we first need to explain the statistical theory that underlies geostatistical interpolation followed by representation of the spatial correlation structure, the basics of geostatistical interpolation (‘kriging’), kriging extensions and spatial stochastic simulation.

The study entitled “sustainability of small farms in coastal ecosystem” was conceived out of consideration of continuous degradation of natural resources- base consisting of soil, water, forest, maintain and animal wealth of the state. With broad objective of investigation into the status of sustainability of agriculture in the state, the specific analysis like socio-economic and farming related situations, perception about sustainability of agriculture in terms of ecological soundness, economic viability, social accessibility, humaneness and adaptability, development of sustainable agriculture index, adoption of sustainable measures, extension components and suggestions and strategy to maintain sustainability were taken into consideration. The organic farming and agri-business opportunities for sustainable agriculture were also studied.

The study was conducted in the state of Odisha covering the districts of Puri, Khurda, Ganjam and Balasore. Altogether four blocks each of one district and eight villages two from each block were selected for the purpose of the study. As much as 200 respondents who were within definition of small and marginal farmers were selected for interview. The respondents randomly so selected were personally interviewed by the author through an interview schedule developed on the basis of objectives stated earlier. The secured data were processed statistically to reveal the relevant findings as per the need of the study. From the relevant findings of the study the conclusions have been summarized below.

The factors affecting sustainability of agriculture under small farm production system in coastal ecosystem of Odisha was studied to ascertain the status of ecological factors and suggest means and ways to improve the situation. The response of the sample farmers and the findings of the study together led to suggest the following measures for up-gradation of ecosystem of the coastal area leading to sustainability of agriculture. Like other states, Odisha is also facing with problems of sustainability of agriculture. Our sample districts namely Puri, Khurda, Ganjam and Balasore which are situated in coastal eco-system of the state has different problems concerning to natural resources like soil, water, forest and animal wealth, etc.

The suggestions to improve the sustainability status of agriculture have been viewed from two important angles. One is adoption of technological measures to preserve natural resources and other is the policy and extension approach to educate people about the necessity of taking such measures for sustainability of agriculture and improve upon.

Transfer of technology is closely associated with generation of technology and its use leading to development. Extension components contain stimulating inputs that would mobilize the practitioners to put recommendations into action. Sustainability in agriculture is very important component of present day agriculture and future aspirations. Keeping these aspects in view an attempt was made to enlist suggestions under extension approaches that would enhance sustainability in agriculture in the area under study. The suggestive extension measures to enhance sustainability in agriculture were studied under components of social structure, cooperation mechanism, evaluation technique and marketing network.

At present producers and consumers are concerned about products and their uses. In farm sector agribusiness is gaining momentum because of marketing of farm produce and its availability to the consumers at affordable rate. Agribusiness thus has been identified as important element in improving farm business. The farm business is otherwise known as agribusiness. Agribusiness is an off farm activity mainly concerned with processing and marketing. It implies that in addition to production of farm produce farmers responsibility increases to process the out-put and its sale in the market for better price. The farmers like to approach different organizations, corporation and regulated markets. Therefore attempt has been made at government level to encourage cultivators to invest in the farm business. In order to increase the efficiency of farmers there is need of keeping the business flexible encouraging farmers to invest in farm and allied sector business.

8.1. Ecological Soundness

Conservation of natural resources essentially means maintenance and improvement of ecological status and not permitting them for further degradation. The natural resources are soil, water, forest, mountain and animal resources. An attempt was made to find out whether the respondents adopt certain recommendations for preservation of these natural resources irrespective of source of advice.

i. Soil Health

Soil is considered as an important ingredient of ecological soundness of an area. Adoption of soil health measures, like adding of organic manures, following of crop rotations, use of balanced chemical fertilizers, green manure and bio- fertilizers are some of the recommendations to maintain soil health. While examining the adoption of these measures, the following results were obtained.