比較農用無人飛行載具與傳統人工方式噴藥和施肥對有機青花菜生產之影響

Abstract

本項田間試驗係在彰化縣埤頭鄉明道大學有機農場進行，比較農用無人載具 (T) 和傳統人工 (CK) 生產管理方式，是否造成有機青花菜在生長和生產上之差異。根據2021 年秋作田間試驗初步結果，所調查之株高、莖長、莖圓周長、葉片數、葉面積及葉面積指數等6 項生長性狀，在此兩種管理方式之間皆未有明顯差別，且葉片、莖及全株的鮮重與乾重亦然。兩方式田間畦面萌生之雜草族群數量，在青花菜生育中期以後即無區別。比較田區四周掛置之黃色 (Y) 與藍色 (B) 黏蟲紙筒的黏捕蟲數，以Y 之捕蟲數大於B 之捕蟲數，而每個Y 平均捕蟲數以CK 田區較高，每個B 平均捕蟲數則以T 田區較高，然而合併Y 與B 之捕蟲數總和，CK 及T 之間並無差異。有機青花菜第1 次採收時，花莖長度平均值在處理之間未有顯著差異，第2 次採收時以T 區大於CK 區，合併兩次採收資料亦以T 區 (23.46 ± 0.54 cm) 大於CK 區 (22.46 ± 0.54 cm)。在採收花蕾球的最大直徑上，第1 次採收的平均值以T 區高於CK 區，第2 次採收兩處理之間未有顯著差異，合併兩次資料後亦無差異。綜合兩次採收之花蕾球鮮重與花莖長度數據進行回歸分析，此兩項性狀在CK 及T 皆未呈現顯著一次或二次回歸關係；但是，綜合兩次採收之花蕾球鮮重與花蕾球最大直徑一次回歸分析結果，則在CK (R2 = 0.415, P < 0.0001) 及T (R2 = 0.229, P < 0.0001) 皆達顯著水準。在公頃花蕾球顆粒數，合併兩次採收數量之平均值，以T 區 (8,047 ± 1,213 no. ha-1) 高於CK 區 (5,506 ± 711 no. ha-1)。在公頃花蕾球鮮重產量上，合併2 次採收計算的平均值，亦以T 區 (3,956 ± 639 kg ha-1) 大於CK 區 (2,531 ± 544 kg ha-1)。進一步分析兩種管理方式採收花蕾球的個別鮮重大小級數百分比分布情形 (分為8 個等級)，CK 區在第1 次採收的花蕾球以重量較小的1–3 級分布較多、第2 次採收則以2–5 級分布較多，T 田區在第1 次採收時以2–5 級分布較多、第2 次採收時則以3–4 級分布較多；合併兩次採收的花蕾球數據，CK 區概以1–5 級分布較多，合計約94.29%，T 區則以2–5 級分布較多，合計達90.50%。顯示在本試驗條件下，相較於傳統人工管理方式，農用無人機管理方式可以生產出較大鮮重的花蕾球。 A field experiment was carried out at an Organic Farm of Mingdao University, Pitou Township, Changhua County. Two field management practices, agricultural unmanned aerial vehicle (or agricultural drone; T) and traditional human labor (CK) for biopesticide spraying and fertilization, were compared to evaluate the differences in the growth and production of organic broccoli. According to the preliminary results of Fall Crop in 2021, there were no significant differences between the two management practices in plant height, stalk length, stalk circumference, leaf number, leaf area, and leaf area index. All traits followed a similar trend as the growth progressed, indicating that the T-managed practice did not cause a significant growth lag in the investigated growth traits relative to those of the CK practice. There was generally no significant difference between the two practices in the fresh weight and dry weight of leaves, stalks, and the whole plant along plant development. Both practices used hand weeding to remove weeds that emerged on the furrows of the field plots. It showed that the furrows with CK practice 19 d after seedling transplanting had a higher weed population. After that, no difference was found between the two practices, and the weed population gradually decreased with the expansion of the broccoli plant size. From the counts of insects caught by yellow (Y) and blue (B) sticky paper traps hanging around the plots of the field, it was shown that counts of insects caught by Y paper traps were greater than that of the B paper traps in the plots of both practices. The average counts of insects caught by each Y paper trap were higher in plots of CK, while the average counts of insects caught by each B paper trap were higher in plots of T. There was no significant difference in the total counts of insects caught by Y plus B paper traps between two management fields. In the first harvest of flower heads (I, 12/17/2021), no difference was found in the average values of the length of flower stalks between practices. In the second harvest (II, 12/24/2021), the average value of the length of flower stalks of T (23.49 ± 0.34 cm) was longer than that of CK (22.20 ± 0.51 cm). The pooled data of two harvests also showed that T (23.46 ± 0.54 cm) had a higher value than CK (22.46 ± 0.54 cm). When the longest diameter of harvested flower heads were compared, the average value of T (19.36 ± 1.11 cm) was higher than that of CK (17.36 ± 0.40 cm) in the first harvest, but not in the second harvest, nor in pooled data of two harvests. By the regression analyses between the fresh weight of flower heads and the length of flower stalks of pooled data, there exists no relationship in both the first- and the second-order regression models. However, the regression between fresh weight of flower heads and longest diameter of flower heads showed a significant linear relationship, in both CK (R2 = 0.415, P < 0.0001) and T (R2 = 0.229, P < 0.0001). With pooled data of two harvests, results indicated that T practice (8,047 ± 1,213 no. ha-1) collected more flower heads than CK practice (5,506 ± 711 no. ha-1), as well as in the yield of fresh flower heads, 3,956 ± 639 kg ha-1 for T and 2,531 ± 544 kg ha-1 for CK. Comparing the percentage distributions of different levels (fresh weights) of flower heads collected in two harvests of both management fields, from level 1 to level 8 (from small to large), it was shown that CK practice collected more flower heads of level 1–3 in the first harvest and level 2–5 in the second harvest. For T practice, it collected more flower heads of levels 2–5 and levels 3–4 in the first and the second harvests, respectively. With pooled data, CK practice collected most the flower heads in levels 1–5, 94.29% in total, while T practice had most flower heads in levels 2–5, a total of 90.50%. As a result, under the conditions of this experiment, compared with the traditional human labor management practice, the agricultural drone management practice may produce flower heads with larger sizes and fresh weights.