Abstract:
This study was aimed at establishing allometric models for estimating LA (Leaf Area) and branch
biomass of eight Coffea arabica genotypes in Mana district of Jimma Zone Oromia Regional State,
South Western Ethiopia. Many Methodologies and instruments have been devised to facilitate
measurement of leaf area and branch biomass. However, these methods are destructive, laborious
and expensive. For modeling leaf area; leaf width, leaf length and leaf area of 1,200 leaves (50 leaves for
each genotype) was measured for model construction and the respective measurements on 960 leaves
were used for model validation. For modeling branch biomass branch diameter of 960 branches (40
branches for each genotype) were measured for model construction and the respective measurements
on 360 branches were used for model validation. Linear measurement was taken from leaves and
branch diameter of eight genotypes of C. arabica, cultivated in field following a randomized complete
blocks design at three altitudes (high, medium and low) were evaluated to identify best option for
input in the models, and to validate the method to estimate the leaf area and branch biomass. Linear
and non-linear models were tested for their accuracy to predict both leaf area and branch biomass of
the eight C. arabica genotypes. The use of linear model resulted in high accuracy for all of the eight
C. arabica genotypes. No significant effect of growing altitude and genotype was obtained among the
slopes of the models for leaf area estimation but there was a significant effect of genotype on the
models for branch biomass estimation. For leaf area estimation one single model was fitted to the
combined data of all genotypes at all altitudes (LA = 0.6434LW). Comparison between observed and
predicted leaf area were made using this model in another independent dataset, conducted for model
validation, exhibited a high degree of correlation (r= 0.98–0.99 P < 0.01).)The over or under
estimation of the leaf area using this model ranges between 0.02 to 1.7% and this model is adequate
to estimate the leaf area for the eight C. arabica genotypes. In the contrary genotype specific models
were developed for each genotype for branch biomass estimation and one general model
(BM=62.059x-2.0532) was developed for all genotypes. Comparison between observed and predicted
branch biomass were made using genotype specific and general models in another dataset, conducted
for model validation, exhibited a high degree of correlation (r= 0.83–0.96, P < 0.01).General models
over or under estimates branch biomass with more than 13 % and less than 38% in three of the coffee
genotypes but the over or under estimation of the genotype specific model is between 0.3 to 7.8%.
Hence genotype specific models can satisfactorily estimate the Branch biomass of the eight C.
arabica genotypes at different altitudes
