Abstract: Frost is one of the major climate disasters in tea production, affecting sustainable development of tea production. Based on the meteorological data of 70 basic weather stations in Zhejiang Province and 25 regional automatic stations in Songyang County, as well as Optimized Consensus Forecast  fine grid numerical prediction products and 30m resolution DEM data, a minimum air temperature model was established by GIS and mixed interpolation method. Taking daily minimum air temperature as tea frost damage warning index, combined with the map of tea plantation extracted from fineresolution ZY-3 satellite remote sensing data, a refined earlywarning method of tea frost disaster was studied. Results showed that the daily minimum air temperature model was of high accuracy with an absolute error of 0.35 ℃ and a relative error of 2.83%. The frost damage warning effectiveness was 8 days in advance and the spatial resolution was 30 m×30 m. Taking the tea frost disaster warning on March 15, 2017 as an example, the preestimated area of affected tea plantation was 2340.0 ha, accounting for 30.6% of the total tea area in the county. The mildly, moderately and severely affected areas were 1939.4, 367.2 and 34.9 hm², which accounted for 25.35%, 4.08% and 0.46% of the county’s total tea area, respectively. The estimation was consistent with the field investigation.

Key words: micro-sprinkling hose irrigation, soil water distribution, hose length, grain yield, dry matter accumulation, wheat