PhD

A project titled "Productivity classification of
soils under rubber (Hevea brasiliensis Muell.
Arg.) in Kerala was taken up in the College
of Horticulture, Vellanikkara during 1996-
1997. The objectives of the study were; (1).
To study the soil taxonomy of selected
prominent soils under rubber in Kerala, (2).
To classify the soils according to the norms
of fertility capability classification, (3). To
attempt an introduction of local modifiers
into the FCC based on specific crop
requirements of rubber plant, (4). To group
the soils by numerical classification
technique using important soil properties, (5).
To test the direct and indirect effects of
relevant soil factors on the growth of hevea
and (6). To identify the spectral signature of
rubber using multi band satellite imagery.
The entire study was conveniently divided
into four experiments.
In the first experiment, six prominent
physiographic units, present in Wayanad,
Malappuram, Thrissur, Kottayam and
Pathanamthitta districts were selected for the
study. In each physiographic unit, three soil
profiles were cut one each from the top, middle
and bottom of the catena of a hill re ascertain
the impact of the slope on the soil properties.
Soils were characterised and ciassited
zccording to the norms of soil taxonomy.
Ir. ~,e second experiment, the soils from the
above soil profile were characterised for
ff;~i;lity capability classification. An attempt
W~; made to introduce a local modifier 'm' to
denote the status of available Mg, using 0.08
crnol kg' in the FC classes identified. In this
experiment the soils studied for taxonomic
classification were subjected to numericai


classification using Euclidean distance
measure.
In the third experiment 14 sites in Thrissur
district were selected for studying the soil-plant
relationships. Soil samples were collected at
bimonthly intervals between November 96 and
September 97 to understand the influence of
weather on the availability of nutrients. Plant
girth and height were recorded simultaneously
while collecting soil samples. Soils were
analysed pH, organic carbon, total nitrogen,
exchangeable Ca, Mg, K, Na, Al and DTPA
extractable Fe, Mn, Cu and Zn. An attempt
was made to relate the dynamics of nutrient
availability to the plant growth.
In the fourth experiment, two rubber estate
areas, one in Thrissur and the other in.
Kottayam districts were selected for studying
the spectral signature of rubber. Satellite data
pertaining to different dates were procured'
studying temporal variations in reflectance.
The reflectance of rubber along with teak and
mixed forest was studied in individual bands,
false colour composite and normalised
difference vegetation index images. A
supervised classification was performed in the
images of both the study areas.
The soil taxonomy indicated that the soils of
all the reaches in the selected physiographic
units belong to Ustic Kandihumults, Ustic
Kanhaplohumults, Typic Kandiustults,
Ustoxic Dystropepts, Typic Kanhaplustults.
Careful observation of the taxonomy of soils
indicated that there are differences in soil
properties caused by the topography resulting
into a catenary sequence of soil development.
However, in some physiographic units, there
were no differences in the taxonomic position

among the reaches though there were known
differences in certain soil properties, which
might have been caused by the slope. The
threshold values used for soil taxonomic
classification were such that the soils with
intragroup distance were not subdivided
further into similar taxonomic 'position in all
the three reaches.
There were differences among reaches of
slope and physiographic units with reference
to fertility capability classification. The
differences in FCC of the same
physiographic units might be mainly due to
differences in topography in addition to other
local variations. Among physiographic units
however, both similarities and differences
were observed. The similarities might be due
to the fact that many of the condition
modifiers were measured only in the surface
soil and similarities in these properties cluster
the soils into one group, though inherently
there could be differences in several other
features. The differences might be due to
variations in slope, parent material or any
other feature, which could influence the
composition of surface soils. Inclusion of a
condition modifier, 'm' could identify three
FC classes with deficiency in available Mg.
In numerical classification of soils it was
seen that all soils belonging to Ustic
Kanhaplohumults were grouped with some
members, however, forming individual
clusters because of intragroup distance.
Different soil taxa viz. Typic Kanhaplustults,
Typic Kandihumults and Ustoxic
Dystropepts formed individual clusters.
However, Typic Kandihumults were merged
to the cluster of Ustic Kanhaplohumults
which might be because of weighted
averaging of soil properties in A and B
horizons. It is seen that similar taxa were
grouped differently in numerical taxonomy
. because of the quantitative differences in soil


165
properties within the same taxon, where in
soil taxonomy certain fixed values of the
attributes are used to classify the soils.
The data on soil variables measured in
different experimental sites at different times
of observations indicated clearly that sites
varied in soil properties. Similarly there were
seasonal changes in soil properties from time
to time. Seasonal changes in plant growth
were seen in all' the experimental sites.
Different rates of growth in the experimental
sites were noticed in response to changes in
weather from a dry season to a wet season,
which also manifested III differential
availability of nutrient elements.
The factor analysis of the data on soil variables
measured at different times of observation
identified two important factors i.e. 'Soil
Reaction Control Factor' and 'P Limitation
Factor' which explained much of the variability
in the data on soil properties. The regression
analysis identified that the growth was
associated with 'P Limitation Factor'. It is
understood that the soil temperature and soil
moisture were the most important parameters
regulating the availability of elements
particularly that of P and K. The limitation
posed by reduced soil moisture, increasing
temperature and DTPA-Mn on the availability
of P was realised. Factor analysis also
highlighted the role of K and Zn in the growth .
The effects of various soil properties on growth
performance were disentangled into direct and
indirect effects in the path analysis. It was
observed in general that the path model could
explain variability in plant growth to a
sufficiently larger extent emphasizing the role
of soil parameters and also highlighting the
interaction effects on plant growth.
While attempting to test the agronomic utility
of soil taxonomy and fertility capability

,


classification, it could be noticed that the soil
taxonomy could give a general order of
ranking based on plant growth among different
taxa. However, it was also noticed that same
taxon was ranked low in some experimental
sites calling for attention while interpreting the
agronomic utility of soil taxonomy. It was
seen earlier that the 'Soil Reaction Control
Factor' and 'P Limitation Factor' influenced
the growth performance of rubber plants.
Soil taxonomy, however, did not contain any
information about the local variations in
these 'factors' and thus could not explain the
variations in plant performance within the
same taxon.
Discrepancies were found while studying the
agronomic utility of the FCC concept also.
Soils deficient in available K and Mg besides
other known limitations were ranked high. It is
felt that efforts should be on to include not


166
only the limited soil variables but also the
variables which limit the availability of the
nutrient elements, to make the FCC system
more meaningful.
~t was clear from the study of spectral
signatures that rubber could easily be separated
from' other vegetation types owing to its
specific spectral behaviour particularly in band
4 (0.77 - 0.86.Jlm). It was also seen that
studies on temporal changes in spectral
behaviour of rubber might help in separating it
from teak as well as mixed forest. It could be
drawn from the data that mean reflectance of
above 60 per cent in band 4 might indicate the
vegetative cover of rubber. Similarly, the
image pertaining to the period when complete
canopy is developed would help in delineating
rubber from teak and mixed forest in these
study areas.