| dc.description.abstract | In general, fertilization practice in sugar cane plantation does not consider spatial variability of soil fertility. The implications of this practice are wasting of fertilizer, decreasing productivity, increasing production cost, decreasing profit, and giving negative impact to the environment. This problem can be solved with application of precision farming approach. Many researches on precision farming, including fertilization activity, have been using many tools like geostatistics, artificial neural network, geographical information system, and decision support system. Nevertheless, those researches just use one tool or combination of two or three tools. This research was done by combining all of the fourth tools. The objectives of this research are: a) analyzing spatial variability of nutrient content of N, P, and K in sugar cane field; b) analyzing spatial variability of productivity in sugar cane field; c) determining N, P, and K requirement at a certain target of yield and sugar content; and d) developing decision support system software to support N, P, and K fertilization based on precision farming approach. The research was conducted from April 2002 until July 2003 in PT Gula Putih Mataram sugar cane plantation, Mataram Udik District, Central Lampung Regency, Lampung Province. The experiment used three treatments of fertilization namely: a) dosage based on the literature reference, b) dosage based on productivity target, and c) uniform dosage. Every block of experimental plot was divided into grids. The determination of fertilization dosage are based on the soil and leaf analysis. Soil and leaf samples were taken with grid center method. Applications of fertilizer were done manually. Fertilizers used were Urea + TSP for basic fertilization and Urea + KCl for second fertilization. The observation of vegetative growth was done on stem population, height of plant, diameter of stem, total number of leaf, and gap percentage. Besides, total number of sugar cane lodge, soil water content, percentage of covering weeds, and level of plant disease attack were observed. Yield is not determined from actual harvest yield but was predicted based on plant population and stem height. Backpro2N, GS+ for Windows, Surfer 8 and ArcView 3.3 were employed for software development. This research showed that precision farming approach in fertilization of N, P, and K could push the spatial variability of nutrient content and productivity. It also showed a moderately low to moderate spatial variability for N and moderately low to moderately high spatial variability for P and K. Land productivity has low to rather low spatial variability. Decision support system that has been built (STRAFERT-PF) can be used to: a) predict yield; b) determine N, P, and K requirement at a certain target of yield and sugar content; c) determine fertilizer dosage; d) analyze spatial variability; e) make field information map; and f) perform cost analysis. The result of this research showed that fertilization with precision farming approach could push the percentage of wasting fertilizer up to 53.47% for Urea and at least 86.96% for TSP; increased productivity 7.6 – 10.7%, and increased profit 1.09%. It also showed that fertilization with precision farming approach could push excessive fertilizer for KCl and increase the plant vegetative growth. | id |
