| 摘要: | 台灣地區近三十年來氣溫增加有明顯增快的現象，除了全球溫室氣體濃度增加造成氣候變遷的影響外，區域人為活動對於環境進行開發，影響土地利用，使地表-大氣間能量收支變化，也會影響局地天氣和氣候的變化。本文試探討長期氣候變遷與土地利用對台灣區域尺度氣象場的影響。利用WRF模式模擬1990-1992年與2010-2012年之春(四月)、秋(十月)兩季，並更換土地利用作為人類活動影響環境之代表，藉此討論土地利用對於長期氣象場的重要性。第一部分，討論大尺度氣象場變化情形，並統計台灣地區近二十年之氣象場變化趨勢。第二部分討論在相同模式設定下土地利用型態對WRF長時間模擬之影響。第三部分討論土地利用與氣候變遷對長期區域氣候模擬的影響。
因為土地利用資料的限制，本研究僅討論近二十年台灣地區土地利用變化所造成的影響，選擇WRF模式中原有的USGS當作1990年代的土地利用代表，並重新處理MODIS衛星資料，使用2001年與2011年資料當作2000年代與2010年代的土地利用資料(命名為MODIS2001與MODSI2011)，以更新整個東亞地區之土地利用。和內政部資料相比，使用MODIS衛星土地利用資料更能真實的代表台灣地區實際土地利用情形，而MODIS2001和MODIS2011只有在零星地區有所差異。在春季(四月)模擬中，USGS的中央山脈部分以MODIS資料取代，將原有的灌溉田取代為森林，以更符合真實情形。
在WRF模式中會有地面風速過高的偏差值存在，本研究使用Jiménez and Dudhia (2012)新增之考慮次網格地形之作用，簡稱topo，研究結果顯示，使用topo可降低風速的偏差值與均方根誤差，改進風速模擬結果。
以中央氣象局地面測站做驗證，比較兩種不同土地利用在2010-2012年十月份的氣象模擬結果顯示，使用MODIS2011的風速模擬結果較接近觀測值。分析模擬的日變化特性，更新土地利用後氣象變數與觀測值較接近，而且土地利用更新對於能量通量影響甚巨，進而影響邊界層高度。更換土地利用會使西半部地區日間溫度增加，風速減低，影響能量通量，夜間由於都市分布，溫度增溫大於日間，在春季與秋季模擬皆有相同的變化趨勢，但由於春季日變化較小，更換土地利用對於氣象場之變化不如秋季模擬結果明顯。由此實驗可知，土地利用對於長期氣象模擬的效應明顯，對於探討模擬長期氣候變遷可為討論的變因之一。
統計台灣地區近二十年氣象場於十月與四月之變化。十月的溫度增加明顯，東北部地區風速增強，全台降雨有增加的現象。四月在東部地區有溫度略微降低、風速增加的現象，而其他地區多為溫度增加、風速減小，四月與十月的雨量變化趨勢明顯不同，四月降雨量有減少的趨勢。
模擬近二十年之氣候變遷，秋季太平洋高壓增強，東北季風減弱，相對的西南風分量使南方富含水氣的氣流與原有的東北季風於台灣東部外海輻合，使東半部地區降水增加；春季方面，近代太平洋高壓北偏，使台灣地區東部外海東風分量增加，使迎風面區域有降水增加，台灣西半部地區背風面降雨減少。溫度方面，十月全台溫度增高，四月則有東半部地區溫度減少、西半部地區溫度增加的情形。絕大部分的模擬變化趨勢與實際觀測之統計變化趨勢一致，因此使用WRF進行台灣地區的氣候變遷模擬實驗有其可行性。
最後討論土地利用與單純氣候變遷，對於近二十年十月與四月之溫度、風速變化的貢獻量。由此結果顯示，土地利用之貢獻量對於長期模擬的影響重大且不可忽視，對於討論局地氣候變遷之議題，土地利用的影響須審慎考慮。由於使用的土地利用資料皆由衛星提供，台灣地區需有實際調查之長期土地利用資料庫，以增加台灣地區長期氣候模擬的可信度。;According to the observation data, the warming is significant over the latest 30 years in Taiwan. Despite of the large scale weather system changed, the changes in land use/land cover characteristics will affect the surface-atmosphere energy budget, and impact on the weather system. In this study, we try to understand the global climate change effect and the local land use change effect on regional climate change over the latest 20 years in Taiwan. The Weather Research and Foresting (WRF) model was applied to simulate the spring (April) and the autumn (October) in 1990-1992 and 2010-2012. To emphasize the human activity affects to the environment, we update the land use data which is derived from the MODIS satellite products. In part one, we analysis the large scale climate change in Asia, and do statistics the observation data in Taiwan over the past 20 years. In part two, we focus on the land use effect under the same model setting in the long term simulation by WRF model on 2010-2012. In part two, we discuss the effect of lands use change on climate in the Taiwan area is important or not. The simulation periods are 1990-1992 and 2010-2012.
Since the long term land use observation data is limited, we use the USGS data from the WRF model as the 1990s’ land use, and we reprocess the land use data from the MODIS satellite as the 2000s’ and 2010s’ land use (the MODIS2001 and the MODIS2010) to update land use in Asia. The new land use data from MODIS is more realistic and can represent the truth land use in recent Taiwan. There are only little difference between MODIS2001 and MODIS 2011. In the spring simulation set, the CRM area change from irrigated land to forest by MODIS data, which is more realistic than the old one.
The WRF model have the high wind speed error. The wind speed correction technique by Jiménez and Dudhia(2012) is used ,which can reduce the bias and RMSE of the wind speed.
Using two different land use data set to simulate the Oct, 2010-2012, the wind speed in MODIS2011 is much closer to the observation data. The characteristic of diurnal cycles are also difference. The update land use can change the energy budget, and then modify the planetary boundary layer height. The updated land use can make the temperature higher and wind speed lower in the daytime over the western part of Taiwan. In the nighttime, the urban island effect is obvious in the urban area. Therefore, the land use data change is really important form the long-term climate simulation.
By the historic observation data, it is much warmer in the October. The wind speed is increasing in the North-East area of Taiwan. The rainfall also increased in October. In April, the temperature is decreasing and the wind speed is increasing in the North-East area of Taiwan, but the other part of Taiwan have the opposite behavior. The rainfall is increasing in April, which is different from October.
The simulation results in spring and autumn over the past 20 years in Taiwan exhibits as below. In October, the tendency of climate change are: weaker Siberia High, stronger south-west wind (weaker north-east wind). In April, the easterly wind component is increased in the eastern coastal of Taiwan. The rainfall increases in the eastern Taiwan (windward side), and decrease in the west part (leeside). The temperature increases all over Taiwan in October, but increase in the east part and decrease in the west part in April. Most of the change tendencies are the same as the observation.
Both contributions of climate change and land use change on the regional meteorological fields cannot be ignored over the past 20 years in Taiwan. The correct land use data is important for the regional weather research model, especially for the climate issue. |
