MODELLING TEA (Camellia (L) O. Kuntze) YIELD USING
SATELLITE DERIVED LAI, LANDUSE AND METEOROLOGICAL
DATA
R.M.S.S. Rajapakse*, Nitin K. Tripathi**, Kiyoshi Honda***
Space Technology Application and Research Program,
Asian Institute of Technology,
P.O. Box 4, Klong Luang, Pathumthani, 12120, Thailand
Tel (66-2)-524-5577 or (66-2)-524-5670
Fax (66-2)-524-5597
Email - stb997175@ait.ac.th
Key Words: LAI, NDVI, TEA YIELD MODELLING
Abstract
Tea is very important cash crop for Sri Lanka as it is the number one foreign exchange earner. Monitoring tea yield is vital from both management of tea plantation and commercial purpose. Tea yield depends on various land use and environment parameters. Remote sensing provides useful information on existing crop condition. In this analysis an attempt has been made to predict tea yield using remote sensing and other key parameters in GIS environment.
Leaf area index (LAI) is one main key factor useful in crop growth models that may be derived from optical remote sensing data. The LAI during the plucking stage is an important state variable in tea yield modeling. The LAI is a major factor determining crop reflectance and is often used in crop reflectance modeling. Therefore, the relationship between tea LAI and optical remote sensing parameter (NDVI) was investigated and this is very important in developing yield prediction models for tea. A model with a logarithmic function was adopted to find the relationship between NDVI and measured LAI.
Satellite derived LAI values and existing spatial, meteorological and agronomic variables with statistical regression analysis and analytical capabilities of GIS were used to develop a model for tea yield estimation. Although different methods were tried to find the best model by using multiple regression, a model develop by using weights of considered variables selected as the suitable model for predicting tea yield.
1.Introduction
Tea (Camellia sinensis (L) O. Kuntze) plantations are very important for Sri Lanka as in terms of land use and employment, tea industry occupies the foremost place in the island. Develop a model to forecast or estimate tea yield is very useful for decision making in tea industry. Remote sensing and GIS technologies has been used during last decade for this purpose for develop yield prediction models for several annual crops like rice, wheat etc.
Therefore, development of a yield prediction model using remote sensing and GIS for tea, a very important perennial crop is pressing need. The lack of previous studies in monitoring tea using remote sensing provided the main *Master student, AIT and Experimental Officer, Tea Research Institute of Sri
Lanka.
**Assistant Professor
***Associate Professor
imputes for this study. LAI plays an important role in both the processes of crop growth and canopy reflectance (Clevers et. al., 1994). Measuring LAI in the field is time consuming. Therefore much relative benefit might be obtained from the estimation of LAI from optical remote sensing data. The normalized difference vegetation index (NDVI) could be considered as a measure of plant productivity (Sellers, 1985). NDVI has been considered a measure of LAI for most of crops (Gong, P. et. al., 1995). Objectives of this study were to develop a model for estimating LAI using optical satellite data and to develop a model for estimating tea yield using LAI, agronomic, landuse and meteorological parameters.
2. Methodology
2.1. Study area
A mid country tea estate situated in Nawalapitiya, Dolosbage planting district of Kandy administrative district in Sri Lanka was selected as the study area. Total land area of the estate is 945ha and 235.25ha is under tea plantation. Balance area is under fuel wood, paddy, vegetables, a tea factory and settlements. Whole estate consists of 50 tea fields respectively. The location of the estate is between 197000mN, 176000mE and 201000mN, 180000mE local coordinates. Several clones of vegetatively propagated tea can be seen in the estate; TRI 2025, TRI 2023, CH13, DN and TRI 2026. All most all seedling tea fields are more than 80 years old. Slope of the estate is less than 53%. Elevation range of the estate varies from 1000m amsl to 1500m amsl. Cumulative annual rainfall of the estate is greater than 2000mm. Average relative humidity of the estate is between 84%-90% and average daily temperature is between 21 0 C to 23 0 C. The base temperature for tea is 13.5 0 C. Therefore this estate has a suitable temperature range through out the year. Daily average sunshine hours of the estate are 5.2 hours.
2.2. Data collection
This study considered both
attribute and spatial data. The attribute data was
collected from tea estate records, meteorological
records and by measuring field-by-field Leaf Area
Index. The spatial data is extracted from
satellite images and existing maps. These spatial
and attribute data were linked within a GIS
database. Existing maps were used to prepare
digital coverages of field boundaries, landuse,
soil boundaries, road and stream network, slope,
elevation and aspects. Using inverse distance
interpolation (IDW) method annual cumulative
precipitation, average daily relative humidity,
average daily sunshine
