小油菜與小白菜節水灌溉之研究

Abstract

臺灣104年農業灌溉用水占總用水量70%，農業灌溉用水中有60%為地面灌溉，然而地面灌溉之水分利用效率不佳，使用水量高於作物所需水量。本研究以小油菜‘青龍’(Brassica rapa ‘Qing Long’)與小白菜‘鳳京’(Brassica rapa var. chinensis ‘Feng-Jing’)為試驗材料，進行持續斷水試驗與斷水復水試驗。持續斷水試驗調查兩作物在停止灌溉後、漸漸進入乾旱逆境時的非破壞性生理特性SPAD、葉綠素螢光FV/FM、NDVI、PRI以及破壞性生理特性地上部之鮮重、乾重、乾物率、葉片水分潛勢與抗氧化酵素活性，並搭配介質泥炭土體積含水量(volumetric water content, VWC)，挑選出適合的非破壞性生理特性作為灌溉指標；斷水復水試驗則調查兩作物在不同供水處理下之地上部鮮重及乾重，並測定泥炭土之VWC，了解兩作物適合之復水值。小油菜試驗重複I中，適合作為灌溉指標的非破壞性生理特性為葉綠素螢光FV/FM，而需採收之泥炭土VWC臨界推薦值為30%；推薦復水之泥炭土VWC為38%。小油菜試驗重複II中，沒有非破壞性生理特性適合作為灌溉指標，採收之泥炭土VWC臨界推薦值為20%；推薦復水之泥炭土VWC為18%。小油菜試驗重複III並無乾旱逆境。小白菜試驗重複I，沒有非破壞性生理特性適合作為灌溉指標，採收之泥炭土VWC臨界推薦值為20%；推薦復水之泥炭土VWC為39%。小白菜試驗重複II，適合作為灌溉指標的非破壞性生理特性為葉綠素螢光FV/FM與NDVI，而採收之泥炭土VWC臨界推薦值為20%；推薦復水之泥炭土VWC為22%。小白菜試驗重複III沒有非破壞性生理特性適合作為灌溉指標，而採收之泥炭土VWC臨界推薦值為20%；推薦復水之泥炭土VWC為18%。綜之，葉綠素螢光FV/FM比其他非破壞性生理特性較適合作為灌溉指標，但仍不夠穩定。於晴天或陰晴參雜時，建議兩作物在泥炭土VWC低至30%之前復水，陰天則建議泥炭土VWC低至20%之前復水。保守的建議為兩作物在泥炭土VWC低至40%之前即復水。亦利用類神經網路預測小白菜產量之分級，成功預測之正確率達31%，若有更多試驗單位將可提升預測產量之正確率，邁向智慧農業。

In Taiwan, 70% of all consumed water was agriculture irrigation water, in which 60% was used in surface irrigation during 2015. However, water use efficiency of surface irrigation is low due to the irrigation water applied is more than plant requirement. In present study, rape ‘Qing Long’ (Brassica rapa) and pak-choi ‘Feng-Jing’ (Brassica rapa var. chinensis) were used as the materials for experiments of (1) water dis-supply and (2) re-water after water dis-supply, for 3 replications. In water dis-supply experiment, non-destructive physiological characters including SPAD, chlorophyll fluorescence FV/FM, NDVI and PRI, and destructive physiological characters including fresh weight, dry weight and dry matter ratio of shoots, leaf water potential and antioxidant enzymes activities, as well as volumetric water content (VWC) of peat medium were investigated after water dis-supply to select suitable character as irrigation indicator. In re-water after water dis-supply experiment, shoot fresh weight and dry weight of 2 crops were investigated under different irrigation treatments. Peat VWC was also measured to select suitable indicator for re-watering. In rape experiment replication I, the suitable non-destructive physiological character is chlorophyll fluorescence FV/FM. The recommended peat VWC for harvest is 30% and for re-watering 38%. In rape experiment replication II, any non-destructive physiological character was not suitable as an irrigation indicator. The recommend peat VWC for harvest is 20% and for re-watering 18%. In rape experiment replication III, the plants did not face drought stress. In pak-choi experiment replication I, no non-destructive physiological character was suitable as an irrigation indicator. The recommended peat VWC for harvest is 20% and for re-watering 39%. In pak-choi experiment replication II, the suitable irrigation indicator for non-destructive physiological character is chlorophyll fluorescence FV/FM and NDVI. The recommended peat VWC for harvest is 20%, and for re-watering 22%. In pak-choi experiment replication III, no non-destructive physiological character was suitable as an irrigation indicator. The recommend peat VWC for harvest is 20%, and for re-watering 18%. In conclusion, chlorophyll fluorescence FV/FM has the potential to be an irrigation indicator, but is still not stable. The recommended peat VWC for irrigation is 30% during sunny days or unstable weather, and 20% during rainy days. Peat VWC never below 40% is conservatively suggested. Artificial neural network was also used to predict classification of pak-choi yield. The accuracy rate is 31%. More experimental units are needed to increase the accuracy of predicting the classification of pak-choi yield and reach intelligence agriculture.