透過系統動力學模型評估光電板覆蓋比對於埤塘生態系統服務之影響：以桃園大崙工作站為例

Abstract

居住於台灣桃園區域早期的居民，為因應農業灌溉用水之需求而建造了許多埤塘，數千座的埤塘形成該市特有之景觀，因此桃園市又被譽為「千塘之鄉」，埤塘的功能除了灌溉用途之外，也具有生態維護、自然調節、養分傳輸與滯留，以及文化價值方面等重要功能。 然而，由於太陽能光電技術的發展，設置光電板的空間已不再侷限於陸地，亦適用於水域環境，因此埤塘的大面積水域空間就有了被利用的動機與經濟利益等誘因。近年來台灣為了追求綠能的發展，推動了許多綠能計畫，而桃園市的埤塘水域空間主要係透過覆蓋「浮動式太陽能光電板」(floating photovoltaics, FPV)建立太陽能發電之可再生能源供應源，並且已形成一定規模之產業型態與供應鏈。 然而，對於埤塘覆蓋太陽能光電板的行為，前後經歷過不少的質疑與抗議，其中不乏居民與環保團體對於埤塘附近區域景觀破壞、生態惡化、水質劣化、鳥類棲息破壞所造成的負面影響具有不少的擔憂，研究學者對於埤塘覆蓋浮動式太陽能光電板所帶來的經濟效益，是否能彌補其所造成副作用的經濟損失，也抱持懷疑的態度。 因此，本研究主要以「系統動力學」(System Dynamics, SD)，透過系統動力模擬專業軟體Vensim，對於桃園區大崙工作站內數個埤塘及灌區，進行系統動力模型建模。本研究之系統動力模型分為「水量交換與變動模型」和「養分傳輸與滯留模型」兩個模型。 水量交換與變動模型主要參考並修改自吳瑞賢等人(2019)與林喬莉(2010)的模式設定，為模擬埤塘和灌區與大圳渠道水源、河水堰水源之間的水量傳輸交換情形。而養分傳輸與滯留模型，主要參考並修改自Li and Yakupitiyage (2003)的半集約式池塘養殖營養動態模型，並針對埤塘系統的特性與水量交換與變動模式串聯，額外考慮了水量交換所導致的埤塘養分輸入與輸出，並同時運行埤塘內的養分傳輸與滯留模擬。

最後將會根據模型模擬結果，評估光電板在不同覆蓋比例(0~100%)之下，對於埤塘光電板發電量、提供灌溉水源、氮與磷的濃度變化與養分滯留、溶氧濃度、植物性浮游生物與異營性成分能量密度、沉積氮與沉積磷、水溫、吳郭魚之生物能量、光電板上所沾黏的鳥類排泄物量與所造成的發電效益損失，以及鳥類的數量等所造成的影響。並粗估在不同光電板的覆蓋比例之下，光電板的發電量、養分滯留、吳郭魚販售、提供灌溉水源所帶來的經濟價值，以探討埤塘光電板的適當設置程度，提供埤塘最佳資源配置之參考。 模擬結果整體而言，當光電板的覆蓋比例越多時，能減少埤塘蒸發損失量，以提高一定程度之埤塘對於灌區的供水，並一定程度減少埤塘與灌區對於大圳、河水水源的依賴。 然而，由於光電板的覆蓋比例提高會降低太陽輻射能量對於埤塘水體的輸入，導致自營性的植物性浮游生物生長受到限制，進一步使異營性成分也隨之下降，因此，兩者之能量密度均會降低，於模擬埤塘中有養殖吳郭魚時，此現象在覆蓋率高於約0.3~0.4左右後更為明顯，且最終吳郭魚所能收成的總生物量也會隨光電板的覆蓋率上升而降低。 植物性浮游生物與異營性成分能量密度隨覆蓋率提高而降低，會導致溶氧量、氮與磷在埤塘系統中的自然面向(植物性浮游生物、異營性成分、吳郭魚、沉積物)中整體傳輸量降低，使全年總沉積氮與磷的量下降，而埤塘水中溶氧濃度、無機氮與無機磷的濃度變動劇烈程度則會趨緩。此外，模擬結果中光電板的覆蓋比例提高，也會導致鳥類的數量減少(但此現象與模式假設鳥類數量會隨光電板覆蓋比例提高而遞減有關)，同時導致鳥類所產生的排泄物總量輸入也減少，且由於光電板的面積增加能夠分攤更多鳥類排泄物的沾粘影響，因此鳥類排泄物造成的光電板發電損失率會隨覆蓋比例提高而下降。 雖然模擬結果中光電板的發電經濟價值遠高出其餘生態系統服務之價值，然受限於研究領域的專業限制，本研究之經濟價值估算並不嚴謹，亦尚未探討利害關係人利益分配的問題，因此不建議以追求光電板的發電為首要考量目標。

The early inhabitants of the Taoyuan region (in Taiwan) constructed numerous agricultural ponds to meet the agricultural irrigation water demands. The large number of agricultural ponds has shaped a unique geographical landscape, earning Taoyuan the nickname “Hometown of a Thousand Ponds”. In addition to their primary function of irrigation, these ponds also serve important roles in ecological maintenance, natural regulation, nutrient transport and retention, as well as holding cultural value. However, with the advancement of solar photovoltaic technology, the space for installing solar panels is no longer limited to land but also extends to aquatic environments. As a result, the vast water areas of the ponds have become attractive for utilization due to economic incentives and potential benefits. In recent years, Taiwan has promoted numerous green energy initiatives to support the development of renewable energy. In Taoyuan, the water spaces of the ponds are primarily used to install “floating photovoltaics” (FPV) for solar power generation, establishing a renewable energy supply source. This has led to the formation of an emerging industry and supply chain. Therefore, this study primarily adopts the System Dynamics (SD) approach and utilizes the professional system dynamics simulation software Vensim to conduct system dynamics modeling for several agricultural irrigation ponds and irrigation zones within the Dalun Workstation area in Taoyuan. The system dynamics model developed in this study consists of two models: the Water Exchange and Fluctuation Model and the Nutrient Transport and Retention Model. The Water Exchange and Fluctuation Model is primarily adapted and modified from the models proposed by Wu Ruixian et al. (2019) and Lin Qiaoli (2010). It simulates the water transfer dynamics between agricultural irrigation ponds, irrigation zones, major canal water sources, and river weir water sources. The Nutrient Transport and Retention Model is mainly adapted and modified from Li and Yakupitiyage’s (2003) semi-intensive pond aquaculture nutrient dynamics model. It is further integrated with the characteristics of the irrigation pond system and the Water Exchange and Fluctuation Model, incorporating additional considerations for nutrient inflows and outflows due to water exchange, while simultaneously simulating the internal nutrient transport and retention dynamics within the pond system. Finally, based on the simulation results, this study evaluates the impact of different FPV coverage ratios (0%–100%) on solar power generation, irrigation water supply, changes in nitrogen and phosphorus concentrations, nutrient retention, dissolved oxygen levels, autotrophic (phytoplankton) and heterotrophic energy density, sediment nitrogen and phosphorus accumulation, water temperature, tilapia biomass, the amount of bird excrement deposited on FPVs and its associated power generation losses, as well as changes in bird populations. Additionally, a rough economic estimation is conducted under different FPV coverage scenarios to assess the economic value of power generation, nutrient retention, tilapia sales, and irrigation water provision, with the aim of determining the optimal FPV deployment level and providing a reference for optimal resource allocation in irrigation ponds. Overall, the simulation results indicate that increasing FPV coverage reduces evaporation losses, thereby improving irrigation water supply and reducing reliance on major canals and river water sources. However, as FPV coverage increases, the reduction in solar radiation input into the pond system restricts the growth of autotrophic phytoplankton, which in turn leads to a decline in heterotrophic energy. Consequently, the energy density of both components decreases. In scenarios where tilapia is farmed in the pond, this phenomenon becomes more pronounced when FPV coverage exceeds approximately 30%–40%, leading to a reduction in the total harvestable biomass of tilapia. The decrease in phytoplankton and heterotrophic energy density results in reduced oxygen, nitrogen, and phosphorus fluxes within the natural pathways of the pond system (including phytoplankton, heterotrophic components, tilapia, and sediments), ultimately leading to a decline in total sediment nitrogen and phosphorus accumulation. Meanwhile, fluctuations in dissolved oxygen, inorganic nitrogen, and inorganic phosphorus concentrations in the pond become less pronounced. Additionally, the simulation results show that increasing FPV coverage leads to a decrease in bird populations (However, this phenomenon is related to the model's assumption that the populations of birds will decrease as the coverage of FPV increases.), thereby reducing the total input of bird excrement. Moreover, as the FPV surface area increases, the deposition of bird excrement becomes more dispersed, effectively diluting its impact on power generation loss, resulting in a lower overall power generation loss rate due to excrement fouling. Although the simulation results indicate that the economic value of FPV-generated electricity far exceeds that of other ecosystem services, the economic estimation in this study is not rigorous, as the primary focus of this research is not economics. Additionally, issues regarding stakeholder benefit distribution have not yet been explored. Therefore, this study does not recommend prioritizing FPV power generation as the primary consideration.