| Numerical calculation of sea waves is an important part of marine economy, marine conservation,and marine safety. For now,Numerical modelling is the most common method of wave calculations. Dr.Donelan and his team developed a new third generation wave model called UMWM(short for University of Miami Wave Model) in 2012. Although UMWM have been applied to simulate daily and hurricane wave field,and also air-ocean-wave coupling research,the research in other areas are rare. The first goal of this paper is attempt to make UMWM apply to global wave simulation,and a global wave model is established after sufficient validations of background errors from altimeter and buoy data,so that to provide a reference for sea wave predicting and related physical process researches. The second goal of this paper is give full play to the advantages of UMWM in air-sea interface flux calculation,by using UMWM to calculate and diagnose the global wind stress and global wave-induced stokes drift.Based on the original parametric source function, the coefficients of parameterizations were improved by contrast experiments and validation with satellite data. To test the model performance after improvement, we used the QuickSCAT/NCEP blended wind field to simulate wave for 9 years, and by using altimeter Jason-1, NDBC buoy data and operational forecast product of the National Marine Environmental Forecasting Center to systematically validate the model results. The results indicate that the UMWM model after improved can simulate the global wave field well,and all validation indexes are close to operational model. By comparing the altimeter data, the results show that: 1) the significant wave height is overestimated when it is small,and is underestimated when it is large. 2) the deviation of simulated wave height exist obviously seasonal and regional differences,the variation tendency on the northern hemisphere is contrary to the southern hemisphere, and the amplitude of variation are similar in low latitude area,but in middle latitude area of northern is twice larger than that of southern. Model error in mid-latitudes area is larger than low latitudes area,and the significant wave height in the west of mainland or swell active area exist an obvious negative deviation region,at the same time the equatorial western Pacific exist a perennial large positive deviation area. By comparing the buoy data,the results indicate that the modeled wave height is basically in accordance with observation in Pacific and Atlantic, and observed dominant wave direction can be accurately and quickly reflected as well,but mean wave period is lower than observation.This paper also analyzes the distribution characteristics of global wave elements in time and space by using 9 years simulating results. By annual average and seasonal average analysis found that the global significant wave height and wind speed both show obvious seasonal change characteristics,and the change rate of northern hemisphere is larger than the southern hemisphere. As affected by the monsoons,the seasonal change laws of Indian Ocean are contrary to other oceans in northern hemisphere. In sum,Global mean wave direction and wind direction present a consistence,but in the central Pacific Ocean where swell active,the angle between wave and wind direction is great,and sometimes even opposite. Mean wave period in the east of ocean is larger than the west, both southern and northern hemispheres have an obvious seasonal change characteristic, but the change rate in northern hemisphere is larger than southern hemisphere.This paper also uses UMWM model calculates and diagnoses the global sea surface wind stress and wave-induced stokes drift. The results illustrate that UMWM modeled wind stress drag coefficient and theoretical results fit well under low wind speed, and under the medium speed show a good consistent with previous observations(Wu,1982;Yelland and Tayler,1996),but appear a little underestimation under the high speed. This indicate that,in the low and middle wind speed,the wind stress observations can be present models and theories explain,but at high speed should be another mechanism has been found. We also found the composition of wind stress at low wind speed is mainly derived from skin stress,when the wave fully develops it turns to form stress. The average annual global wind stress near the equator is small,and near the westerlies is much larger,global wind stress value is between 0.03 ~ 0.35 N/m2. Meanwhile,the global wind stress has an obvious seasonal change characteristic,the maximum value of summer appeared in the southern Indian Ocean,while the maximum value of winter appeared in northern Atlantic,and the change rate in northern hemisphere is larger than southern hemisphere. The average annual global surface stokes drift near the equator and 30°latitude area is small,while the westerlies and equator outside area is larger,most part of global surface stokes drift velocity is between 0.01~0.18m/s. The average annual global stokes drift depth near the southern hemisphere westerlies waters is the largest,and within 30°S increases with latitude,while the global drift depth value is between 1 ~ 4m. Drift velocity and depth has obvious seasonal variation characteristics,for the northern hemisphere except the Indian Ocean, both show the weakest in the summer,then gradually increase,and come to the peak in winter(while the southern hemisphere on the contrary). Global Stokes drift transport value over the annual average is between 0.01 ~ 0.68 m2/s,and its space-time distribution characteristics similar to drift velocity.There are three innovations in this paper. Firstly,an improved source function coefficients apply to UMWM model is proposed.Secondly,comprehensive performance verification on the improved model is carried out,and a comprehensive understanding of the structure and space-time distribution of model error is given. Finally,based on the UMWM model,the diagnosis analysis work of wind stress and stokes drift on the air-sea interface is carried out. |
PDF Full Text Request