Distributed Modeling Of Monthly Mean Temperature Parameter Optimization Combined With Remote Sensed Data

Based on previous researches on temperature model, parameters optimization processes were achieved by clearify and remote sensed uneven distribution features. Aft-parameter optimization temperature model was built by using conventional meteorology station observation data, remote sensed data, etc. In order to accomplish the goal of monthly average temperature distributed assimulation. Results of 2001 monthly temperature characteristic of 1km×1km spatial resolution were also showed in this paper.Contents of the study and related results are as follows:1) Based on physical empirical statistical model theories, problems of clearify of long wave effective radiation parameter was solved, with better physical meanings. Compared the factor of invisible with clear form, later has much definite physical meaning than former. Temperature model with clearify parameter reaches the annual MABE of 0.82℃, declined by 0.16℃ than invisible one. Model’s MABE is inproved after parameter optimization, with much more stable and reliable results.2) MODIS landsurface emissivity data were used to synthetise 2001 monthly landsurface emissivity of China. In order to slove the problem of uneven distribution of landsurface parameter in relative computational issues. Results show that:Area ratios of different emissivity value intervals change respectivly with seasons. High value area ratio is higer in winter while low in summer. Middle value area ratio is higer in summer while low in winter. Low value area ratio changes with less obvious character. Emissivity shows a "Shuangfeng double Valley" feature when aspect changes. Southeast and northwest values are higer, while south and north values are lower. Emissivity shows an obvious regional distribution when altitude increases. Under 6000m, distribution characteristics of emissivity with respect to aspect is similar to that of NDVI, while above 6000m differences of emissivity and NDVI distribution characteristics reflect bare soil and ice covered land features.3) Long wave effective radiation of rugged terrain in 2001 (both month total amount and year total amount) were calculated by considering horizontal effective radiation, terrain openness, etc. Results shows that:annual MABE decreased by 5.07MJ/m2, with the decrease of annual MRABE of 0.55 percentage point, considered the distribution of emissivity rather than a constant 0.95, simulated results are reliable.4) Monthly temperature in 2001 of rugged terrain over China region was calculated by considering altitude, total sun radiation over rugged terrain and long wave effective radiation. Using nearby pixels error analysis method, results show that:monthly temperature MABE reaches lowest in June 0.19℃, annual temperature MABE is 0.24 ℃. Compared with 3 temperature models, results show that:IDW model elegantly described temperature macroscopic distribution, but with worse local discriptions and unstable error change. Classical terrain climate adjust model has better physical meanings and accuracy than IDW model. Optimized model owes much better physical parameters, and the accuracy increases a bit than the Classical model, considerd parameter clarify process.5) Local analysis shows:temperature changes with month maitaines a relative parallel line under different land height. Shows regular pattern of temperature declines with height.45°N, Jan and Oct, temoperature anomaly is positive in sunny aspect, while negative in shadowing one. The feature changes conversly in July. Differences of anomaly in each aspect are much less in Apr. 30°N, with the increase of elevation and slop, differece of temperature anomaly in sunny aspect and shadowing ones increses in Jan. While become less obvious in July. Sawtooth features enbody the influence of terrain openness. In warmer season, temperature lapse rate is heigher than that in cold season. Lapse rate in south aspect is heigher than north in Jan. Differences of lapse rate in each aspect are much less in Jul. Temperature decreases with elecation shows weak differences in Apr and Oct under 4 aspects.Compare to previous studies, innovations of this paper can be summarized as follows:1) Influencing factor such as clearified long wave effective radiation was incorporated into monthly temperature model, to replace former invisible one(the algebra combination of percentage of sun shine and water vapor), in order to improve physical property of the model.2) Based on emissivity remote sensed data, band synthesis method was used to get monthly broad band emissivity. Not only was the problem of uneven spatial distribution of emissivity solved, but also the model physical properties and simulation accuracy has been improved.3) Elevation, total sun radiation under rugged terrain, as well as long wave effective radiation under rugged terrain were incorporated into temperature distributed model. Simulation of aft-parameter optimization distributed model was accomplished. Compared to privious studies, temperature model parameters have been physically and accurately improved.