Geoscience Reference
In-Depth Information
namely, biological data, soil and crop data and atmospheric
measurements. Owing to the complexity and multiplic-
ity of the effects of environmental factors on agricultural
production, it is necessary to use statistical techniques to
detect the interactions of these factors and their practical
consequences. Based on the objectives and interests of the
research, appropriate statistical models starting from mul-
tiple regression analysis to logistic regression to time series
modelling have to be employed. The main objective of this
chapter is to provide the basic idea behind some statistical
tools that can be successfully employed to study the effect
of climatic factors on crop production.
4.1
Introduction
Statistics has two major components: descriptive statistics and
inferential statistics. Descriptive statistics gives numerical and
graphical procedures to summarise a collection of data in a clear
and understandable way whereas inferential statistics provides
procedures to draw inferences about a population from a sample.
Descriptive statistics are used to describe the basic features of the
data in a study. They provide simple summaries about the sample
and the measures. Together with simple graphics analysis, they
form the basis of virtually every quantitative analysis of data.
Descriptive statistics are used to present quantitative descriptions
in a manageable form. In a research study, there may be lots of
measures. Descriptive statistics help us to simplify large amounts
of data in a sensible way. Using the numerical approach one might
compute statistics such as the mean and standard deviation which
are precise and objective. Graphical methods are more suited
for identifying patterns or trends in the data. The numerical and
graphical approaches complement each other. Some of the graphi-
cal approaches are
Box plot
, which is an excellent tool for convey-
ing location and variation information in datasets, particularly for
detecting and illustrating location and variation changes between
different groups of data;
scatter plot
, which reveals relationships
or association between two variables, such relationships manifest
themselves by any non-random structure in the plot;
probability
plot
, which is a graphical technique for assessing whether or not a
dataset follows a given distribution such as the normal;
histogram
and
stem-and-leaf plot
, which is a display that organises data to
show its shape and distribution and so on.
Rainfall is an important climatic variable that imposes crop
production risks, especially on rain-fed subsistence cultivation
