無線感測網路於森林物候監測應用之研究

Abstract

在物候監測研究上，利用科學的方法收集微氣候與植物生長的相關資料，轉化成有用的知識以作最正確的判斷和評估。台灣物候研究的觀察，以往由於野外觀察不易、觀察間隔較長及遠距觀察，且缺乏詳盡且連續的環境資料來確認物候變化的機制，應用無線感測網路和攝影機在物候的監測上，可方便吾人觀察到較以前傳統方式不容易看的資訊，能對於物候能有更進一步的認識。本研究在奧萬大森林遊樂區和阿里山森林遊樂區佈放無線感測網路和架設網路攝影機，監測楓樹林葉變色和吉野櫻花花期的物候現象。在楓樹林葉變色方面，利用影像的RGB數值來檢定葉變色與否，再與溫度的物候模式結合，讓我們可以了解葉變色率與溫度之間的關係。其結果在10月1日過後經累積低溫(<16℃)環境累積和達1020 degree-hrs後，楓樹林的葉變色率則有可能會達到100%。另在阿里山吉野櫻開花的監測上收集自2007年1月至2007年4月的溫度及吉野櫻的影像資料，利用溫度熱量累積的物候模式(degree-hrs model or degree-days model)來推估建立開花的模式。研究結果顯示無線感測網路資料回傳率已達到80%以上，在櫻花花期監測上可知2007的開花日期為2007/3/16，盛開日期為2007/3/29。發現阿里山吉野櫻花開花所需的熱量累積和為8226(模式一) or 4527(模式二) degree-hrs，達盛開日期則需10300(模式一) or 5953(模式二) degree-hrs。無線感測網路和網路攝影機在於物候的監測上提供了方便性和實用性，儀器所偵測到的資料透過網路大量、密集、快速的傳送到實驗室，讓研究者可以迅速的掌握資訊。此應用打破傳統的資料蒐集方式，已成為一個新的研究方法論，尤其在生態研究、資源管理與環境監測上提供了嶄新的運用。

In the phenological monitoring, scientific methods are expected to compile the data of microclimate and vegetation growth, which would be processed into useful information. The researchers could, therefore, make right decision, predict, and assessment in the phenology. The phenological research in Taiwan is not easy to implement, such as long term succession field inventory and observation, and it lacks of detailed and continuous data to confirm the mechanism of the phenological changes. Consequently, this research applied the new advanced technology, the wireless sensor network (WSN) and web-camera, in the process of phenological monitoring, providing more information than traditional method. The study established wireless sensor network and web camera in Auwanda and Alishan Forest Recreation Area to monitor the phonological phenomenon of Liquidambar formosana Hance and Prunus yedoensis cv. Yedcensis relatively. The leaf coloration of Liquidambar formosana Hance is classified by the RGB value of monitoring images. Then, it compiled the data of microclimate to understand the relationship of leaf coloration and temperature. The results showed that the cooling hours needed is 1020 degree-hrs for reaching 100% of the leaf coloration from October 1, 2006. For the blossoming phenology, the temperature was recorded every 10 minutes and blossoming image of Prunus yedoensis cv. Yedcensis were taken from January 1, 2007 to April 15, 2007. According, the phenological model (degree-hrs model and degree-days model) of blossoming were established. The results showed that the feedback rate of WSN was 80% during the studying period. Although the rate seemed to be low, the data was enough to detect that the flowering date was on March 16 and full-flowering date was on March, 29. Meanwhile, the accumulating heat needed for the flowering was 8226(model 1) or 4527(model 2) degree-hrs and full-flowering was 10300(model 1) or 5953(model 2) degree-hrs. The WSN and Web-camera provided convenience and practical method for the phenological monitoring. The device transferred the mass, fast and timely data to the lab through the internet. The researcher could keep the latest information. This new advance application broke through the traditional methods and would become a new research method gradually in ecological research, resource management and environmental monitoring.