地面高光譜系統之建置：評估 NIRvP 作為 SIF 代理指標之環境穩定性

Abstract

太陽誘導葉綠素螢光（sun-induced chlorophyll fluorescence, SIF）被認為是目前最適合作為光合作用替代指標的遙測訊號之一。然而，SIF 的反演需要仰賴次奈米光譜解析度的進階技術，使其量測與應用具有相當挑戰性。相較之下，將近紅外植被反射率（NIRv）與光合作用有效輻射（PAR）相乘所形成的指標，即NIRvP，則具有較為直接且簡便的計算方式。然而，SIF與NIRvP之間的相關性是否能在不同氣象條件下維持穩定，目前仍缺乏充分探討。 因此，本研究建立一套地面雙光譜儀系統，能夠在台灣高山茶園中同步量測下行輻射與上行輻射。該研究場域位於雲霧頻繁的山區，快速變化的霧與雲層動態常造成微氣候的大幅波動。本系統設計目的在於反演SIF，並利用同步取得的太陽輻射資料計算NIRvP。系統配置兩台QE-Pro光譜儀，光譜範圍涵蓋640–804 nm。搭配 餘弦校正器與25.4°視野的裸光纖，分別取得半球式下行輻射與圓錐式上行輻射。PAR則由鄰近氣象站量測。自2025/11/15日至 2026/05/23日，本研究於日內時段（07:00–17:00）以五分鐘間隔連續取得SIF與NIRvP資料後，進一步彙整為30分鐘尺度資料（n = 1573）。 在所有30分鐘觀測資料中，SIF與NIRvP呈現高度線性關係（R² = 0.874），顯示NIRvP能有效捕捉SIF中受入射輻射與冠層結構控制的主要變異。然而，兩者關係並非在時間上完全穩定。局部R²通常在上午至中午時段較高，並於清晨與午後較低；同時，在春季午後，觀測到的SIF逐漸低於整體NIRvP–SIF迴歸關係所預測的數值。此方向性偏差在太陽位於冠層西南至西側、且太陽天頂角相對較高時最為明顯。在所檢驗的氣象變數中，大氣溫度與SIF–NIRvP關係穩定性的變化有最明確的關聯。 本研究結果顯示，NIRvP是SIF的有效替代指標，但其近似 SIF 的能力會在特定日內與季節條件下減弱。因此，未來若以NIRvP近似SIF變異，應同時考量太陽幾何與季節變化所造成的系統性影響。

Sun-induced chlorophyll fluorescence (SIF) is arguably the most suitable surrogate for photosynthesis. However, it requires advanced techniques to derive the metric at sub-nanometer resolution, making the task challenging. A suitable proxy, near-infrared reflectance of vegetation (NIRv) multiplied by photosynthetically active radiation (PAR) (hereafter NIRvP), on the other hand, is straightforward. However, the stability of the correlation between SIF and NIRvP under different meteorological conditions has not been investigated. Therefore, in this study, we developed a ground-based dual-spectrometer system capable of simultaneously measuring downwelling irradiance and upwelling radiance in a montane tea plantation in Taiwan, where rapid fog and cloud dynamics frequently cause large fluctuations in microclimate. The system was designed to retrieve SIF and simultaneously calculate NIRvP using concurrent solar radiation data. The system was equipped with two spectrometers (QE-Pro) covering a spectral range of 640–804 nm. Cosine corrector and bare fibers with a 25.4° field of view were attached to in-house light collectors to acquire hemispherical downwelling irradiance and conical upwelling radiance. PAR was acquired by a nearby meteorological station. Diurnal (07:00–17:00) SIF and NIRvP were collected at five-minute intervals from 2025/11/15 to 2026/05/23, and aggregated to 30-min intervals(n=1573). Across all 30-min observations, SIF was strongly related to NIRvP (R² = 0.874), confirming that NIRvP captured most of the variation in SIF driven by incident radiation and canopy structure. However, this relationship was not temporally uniform. Local R² was generally higher from morning to midday and declined during early morning and late afternoon, while observed SIF increasingly fell below values predicted by the overall NIRvP–SIF regression during spring afternoons. This directional bias was most evident when the sun was located southwest to west of the canopy and at relatively high solar zenith angles. Among the examined meteorological variables, air temperature showed the clearest association with changes in relationship stability. These findings demonstrate that NIRvP is an effective proxy for SIF, but its ability to represent SIF weakens under specific diurnal and seasonal conditions. Consequently, studies using NIRvP to characterize SIF variability should account for solar geometry and seasonal changes.