巴伦支海二甲基硫模型参数校准的进一步探讨
发布时间：2018-11-21 09:16
【摘要】：近几十年来,由于大量的二氧化碳等温室气体排放到大气中,使北极的海表温度升高,海冰融化加速.海洋植物的活动产生的二甲基硫(DMS)排放到大气中通过一系列变化会减弱温室效应.本文研究的是巴伦支海(30°E-40°E,70°N-80°N),提取该区域2003-2009年的环境数据——冰盖(ICE),风速(WIND),云盖(CLOUD),海表温度(SST),叶绿素(CHL),混合层深度(MLD),MLD数据是通过北极的CTD数据计算得到的;2003-2005年的DMS的观测数据是基于研究区域1998-2002年的历史数据用BP神经网络时间序列模型预测得到的,并与用Simo提供的公式计算的DMS数据进行比较;用遗传算法校准DMS模型中的参数,得到该区域的最优参数后,对这些参数进行灵敏度检验;最后用CMIP5(5th Coupled Model Intercomparison Project)试验中21×CO和24×CO的数据,预测和比较DMS海气通量的变化.本文得到的主要结论如下:从2004年开始冰盖整体都有所下降,在2004-2007年的8月到9月,冰盖均降到0%.在南部区域(30°E-40°E,70°N-75°N)的覆盖率整年都很小,2005-2009年几乎达到无冰区域.云盖的月平均值一般都在80%以上,只有2004年的7月云盖低至75.4%.风速通常在7月达到最小值,约5m/s左右,冬季最大风速达12.5m/s.SST最大值一般在8月达到,在这7年SST有下降的趋势,2003年SST最大值为8.2℃,2009年最大值只有5.9℃.2009年CHL的峰值最大,约4mg/m3.MLD在秋季是最浅的,夏季次之,冬季较深,在1月和3月出现了“双峰”现象.用BP神经网络预测得到的DMS的峰值在2004年最大,为11.77--12dumolm.与用Simo公式计算的DMS作为观测值的校准结果相比,用BP神经网络预测的数据作为观测值的校准结果较好.CHL峰值会先于DMS峰值8-16天,DMS海气通量(DMS-Flux)和传播速度(Kw)的峰值是在同一时间段达到的,浮游动物的峰值一般在7月达到.灵敏度检验表明:灵敏度最大的参数是代表浮游植物S(DMSP):N比例的γ,其次是代表细菌对DMS消耗率的28k.在24×CO情况下,SST、WIND、MLD相对于在21×CO下,分别增加了233.684%、22.744%、28.232%,ICE覆盖率下降了11.986%,CLOUD变化不明显,只下降了1.322%.在24×CO情况下,DMS海气通量在有冰和无冰的情况下增长率分别为分别为148.840%、249.679%.本研究得到的结果对极地的生态环境与气候的研究具有很好的参考价值.
[Abstract]:In recent decades, sea ice melting has accelerated due to large amounts of greenhouse gases, such as carbon dioxide, released into the atmosphere, causing the Arctic sea surface temperature to rise. The release of dimethyl sulfur (DMS) from the activity of marine plants into the atmosphere weakens Greenhouse Effect through a series of changes. In this paper, the environmental data of the Barents Sea (30 掳E-40 掳Elae 70 掳N-80 掳N),) from 2003 to 2009 are studied. The (ICE), wind speed of the ice sheet, (WIND), cloud cover (CLOUD), sea surface temperature (SST), chlorophyll (CHL), is studied in this paper. The mixing layer depth (MLD), MLD data are calculated by the Arctic CTD data. The observed data of DMS from 2003 to 2005 were predicted by using BP neural network time series model based on the historical data from 1998 to 2002 in the study area, and compared with the DMS data calculated by the formula provided by Simo. The parameters in the DMS model are calibrated by genetic algorithm, the optimal parameters in the region are obtained, and the sensitivity of these parameters is tested. Finally, the variation of sea and air fluxes of DMS is predicted and compared with the data of 21 脳 CO and 24 脳 CO in CMIP5 (5th Coupled Model Intercomparison Project) test). The main conclusions obtained in this paper are as follows: the whole ice sheet has declined since 2004, and from August to September of 2004-2007, the ice sheet has dropped to zero. The coverage in the southern region (30 掳E-40 掳EN 70 掳N-75 掳N) was very small throughout the year, and almost reached the ice-free area in 2005-2009. The monthly average of cloud cover is generally above 80 percent, only to reach 75.4 percent in July 2004. The wind speed usually reaches the minimum value in July, about 5m/s, and the maximum wind speed reaches the maximum value of 12.5m/s.SST in August in winter. In these 7 years, the maximum value of SST is 8.2 鈩，
本文编号：2346535
[Abstract]:In recent decades, sea ice melting has accelerated due to large amounts of greenhouse gases, such as carbon dioxide, released into the atmosphere, causing the Arctic sea surface temperature to rise. The release of dimethyl sulfur (DMS) from the activity of marine plants into the atmosphere weakens Greenhouse Effect through a series of changes. In this paper, the environmental data of the Barents Sea (30 掳E-40 掳Elae 70 掳N-80 掳N),) from 2003 to 2009 are studied. The (ICE), wind speed of the ice sheet, (WIND), cloud cover (CLOUD), sea surface temperature (SST), chlorophyll (CHL), is studied in this paper. The mixing layer depth (MLD), MLD data are calculated by the Arctic CTD data. The observed data of DMS from 2003 to 2005 were predicted by using BP neural network time series model based on the historical data from 1998 to 2002 in the study area, and compared with the DMS data calculated by the formula provided by Simo. The parameters in the DMS model are calibrated by genetic algorithm, the optimal parameters in the region are obtained, and the sensitivity of these parameters is tested. Finally, the variation of sea and air fluxes of DMS is predicted and compared with the data of 21 脳 CO and 24 脳 CO in CMIP5 (5th Coupled Model Intercomparison Project) test). The main conclusions obtained in this paper are as follows: the whole ice sheet has declined since 2004, and from August to September of 2004-2007, the ice sheet has dropped to zero. The coverage in the southern region (30 掳E-40 掳EN 70 掳N-75 掳N) was very small throughout the year, and almost reached the ice-free area in 2005-2009. The monthly average of cloud cover is generally above 80 percent, only to reach 75.4 percent in July 2004. The wind speed usually reaches the minimum value in July, about 5m/s, and the maximum wind speed reaches the maximum value of 12.5m/s.SST in August in winter. In these 7 years, the maximum value of SST is 8.2 鈩，
本文编号：2346535
本文链接：https://www.wllwen.com/kejilunwen/haiyang/2346535.html
