Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/27541
標題: 田間尺度水分平衡模式在土壤水分模擬上之研究
Study of Applying Hydrologic Budget Model on a Field Scale Simulation of Soil Moisture
作者: 向為民
Hsiang, Wei-min
關鍵字: Hydrologic budget model
土壤水分收支模式
Field scale
田間尺度
出版社: 土壤環境科學系
摘要: 本研究針對目前台灣水資源的管理趨勢及農業部門用水的限制,在現今農田水利會持續執行渠道灌溉的條件下,透過簡單的動態土壤水分平衡模式,掌握田區土壤水分含量即時的變化情形,做為灌區水源調配之依據,以期能夠以符合經濟效益的方法節用水資源,使台灣農業的生產朝向有利環境保育的正面方向發展。模式需要的氣象資料僅限於雨量及蒸發量。土壤及作物相關的模式參數,則利用田間資料反求。同時嘗試馬可夫一階模式,利用其可以藉由三維資料的掌握化為各土層在平面上的變異及在垂直方向的相關。達成以一維的模式來代表田間土壤水分的三維分佈特性。 首先,本研究特從文獻著手,對作物的需水特性、理論模式的發展,以及相關的土壤物理性質,做一回顧,作為本論文之學理研究的發展依據;其次從一般較少處理的田間坵塊尺度,研究土壤水分含量的變異特性,特點在於跨出以定點研究土壤水分管理的傳統思維方式,建立以田間園區坵塊為單位的水分管理方向。這一尺度的建立預期可對區域性大尺度的水資源管理,提供發展基礎。 水分收支平衡模式建立在土壤水分含量隨時間變化量等於水分進出根域,即地表以下60公分深土體,其中水量的差額,入滲水量之計算是以農地計劃排水量為基礎,由降雨量減去逕流量求得。水分在土壤中的傳輸以及根域受毛管支持水的輸水補給,本文認為是與根域土壤含水量(W)呈線性相關,即毛管輸水補給量G=Ka(W-Wfc),Ka為導水速率常數,Wfc為當地狀況條件下根域土壤之田間容水量。農地蒸發散量ET可以由作物係數Kc與潛勢蒸發量ET0求得,即ET=Kc×ET0。Kc隨時間變化之函數為作物係數曲線;玉米生育第一階段為葉面積指數<1或作物覆蓋率未達10%時,此時期Kc值為最小常數;葉面積指數≧3時或作物覆蓋率達70-80%以上時為玉米生育第三階段,此時期Kc為最大常數值。玉米生育劃分為四個階段,其作物係數均以線性方程式來表示。水分收支平衡模式,需要雨量、蒸發量資料,以及Ka、Wfc參數。Ka及Wfc參數值可以由作物生育第一階段時期,利用田間資料求取,惟模式應用時需考慮參數求取時之土壤含水量範圍。本文利用嘉南農田水利會灌溉試驗中心之77年春作,玉米灌溉試驗的兩個不同灌溉時間的處理,所得到的田間土壤水分含量變化的資料,進行模式參數之求取及玉米田土壤水分含量之模擬研究。 結果顯示,在模式建立與驗證上,利用玉米生育初期,0-26天葉面積指數小於1 (LAI<1)的田間調查資料可以順利取得模式的三個參數,分別是Kc,作物係數參數;Ka,土壤導水參數;以及Wfc,係當時環境下之田間容水量。利用參數,配合雨量、灌溉量、蒸發量、作物係數曲線,則可以模擬不同處理狀況下玉米全生育期之土壤水分含量。求得調查值與推估值的直線相關,A及B兩處理其調查與推估值直線迴歸的相關係數R2分別為0.94與0.92均達顯著水準。研究工作顯示,模式可以有效模擬的土壤水分範圍,自93mm至199mm,而且需要的參數極少,參數之取得極為方便且經濟,是可以立即應用在本省耕地上,做為水分管理之工具。 馬可夫一階模式(Marcov Lag-one model)可以考慮三維資料中垂直方向之相關性與平面中存在的變異情形,在本文中則是指各土層水分含量在平面分佈上的平均值與標準偏差,以及上下相鄰土層的直線相關係數。在模擬時,由隨機變數的函數來掌握平面的統計特性。 馬可夫一階模式之測試,乃利用農委會農試所16號田83年秋作玉米田土壤水分含量調查資料。期望時間序列模式得以掌握田間土壤水分之分布及變異特性。結果顯示在降雨的狀況下,0.35公頃農地土壤水分含量的變異係數最高僅達15%,然而經過局部區域未潤濕之灌溉處理後,可將田間土壤含水量的變異係數升高至40%,再逐漸降為20%上下。灌溉處理以後,相鄰土層土壤水分含量的直線關係趨向明顯,直線迴歸係數在0.9上下。 綜合全生育期土壤水分含量,0-15與15-30公分土層的土壤含水量間存在直線相關,R2值為0.7388;然而15-30與30-60公分土層土壤含水量直線相關的R2值則為0.7406。利用模式產生之資料仍保有原土層之族群特性,據此可以推論在大尺度狀況下,土壤水分含量變異係數夠大時,馬可夫一階模式配合蒙地卡羅方法可掌握田間土壤水分含量之分佈及變異特性。 由研究工作所得結果可以知道 (1) 由氣象及土壤水文資料,利用簡單的土壤水分收支模式,可以即時地預測根域土壤水分含量。 (2) 馬可夫一維模式可以掌握田間土壤水分三維特性的變化情形。
The purpose of this study is to establish a hydrological budget model for the simulation of variance of soil moisture in southwest area of Taiwan. Parameters of climate data required are limited only to precipitation and evaporation, while those of the correlation of soil and crop are derived from field survey and simple statistics analysis. Meanwhile, attempt is made using Marcov Lag-one Model, that could well handle three- dimensional data of the horizontal moisture variance in soil layers and their vertical correlation, to formulate an one-dimensional model to present the three- dimensional distribution of moisture in a specific field soil. The hydrological budget model is based on the assumption that moisture content varying with time is equivalent to the water deficit(due to flow in and out)of root zone(60cm depth in a solum), while on the basis of planned drainage amount of water in field ,the amount of infiltration water is derived from the amount of precipitation minus run-off. In areas of southwest Taiwan, water movement in a solum and water supply to root zone by capillary water are observed linearly correlated with soil moisture content(W), i.e., G= Ka (W-Wfc), Ka indicating hydraulic conductivity constant and Wfc representing field capacity in situ. Field evaporation(ET)is obtained from crop coefficient(Kc)and potential evaporation(ET0), i.e., ET = Kc * ET0. Kc is a function according to change of time and defined as the crop coefficient curve, merged by four linear equations expressing crop coefficients of four commonly divided growth stages of corn. When at growth stage 1st, with leaf area index (LAI)smaller than 1 or cover rate of canopy still less than 10%, Kc is the value of the minimum constant. While at growth stage 3rd, with leaf area index larger than3 or cover rate of canopy attaining 70-80%, the maximum constant of Kc value is obtained. Essential for a hydrological budget model are parameters of Ka and Wfc, and data of precipitation and evaporation. Ka and Wfc can be acquired from field data at any growth stage as long as the range of soil moisture content is carefully considered. To study the simulation of soil moisture and secure the parameters for model, an irrigation timing experiment of spring crop of corn(draught period)was conducted in Chia-Nan area. The results indicate parameters of Kc, Ka and Wfc are readily available from field survey of the initial growth stage of corn(0-26 day, LAI<1)in Hsueh-Chia for the formulation and validation of model. Integrating parameters with data of precipitation, irrigation, evaporation and the crop coefficient curve, the soil moisture content of different treatments during the entire growth season can be simulated to get the linear relation between observed and simulated data. The correlation coefficients R2 of treatment A and B, 0.94 and 0.92, respectively, are both extremely significant. Analysis of the hydrological budget shows the precipitation of 273 mm in 1988 spring crop is sufficient to meet the demand of evaporation, which could even be 1.1 times its potential evaporation at the culmination of growth. The Marcov Lag-one Model enables consideration of three- dimensional data of the horizontal moisture variance in soil layers and their vertical correlation. The mean value and standard deviation of the horizontal distribution of soil moisture content in every soil layer are included and considered in this text. Furthermore, the coefficients of linear relation between neighbor layers are also calculated in the study. Random variables or numbers handle the characteristics of horizontal statistics, when simulating. Validation of the Marcov Lag-one Model is carried out with soil survey of field No.16 in Taiwan Agricultural Research Institute. It indicated under conditions of rain and irrigation the maximum variance coefficient of soil moisture content in field of 0.5 hectare was 15%, while that under conditions of partially wet and partially remain drought could reach over 40%. Through assessment of soil moisture content during the entire growing season, linear correlation was found between layers of 0-15 and 15-30 cm, and also between layers of 15-30 and 30-60 cm, with coefficients R2 of 0.7388 and 0.7406, respectively. Data generated from the Marcov Lag-one Model have advantages of maintaining population features of original soil layers and existence of linear relation between two neighbor layers. Therefore, on a large field scale, the Marcov Lag-one Model combined with Monte Carlo Method can be expected to handle the variance and distribution of soil moisture content in field, as long as the variance coefficient is large enough to locate the correlation between every two neighbor layers. Readjustment use of field land is quite crucial to sustainable agricultural development in Taiwan. It is convinced that facilities design of irrigation and drainage, fulfilling requirements for environmental protection, once introduced will spark great change in current environment of soil water. Long-term fallow is not only for ecology reason but contributes to more benefit of conservation of water resource than that of paddy. Quantitized assessment of such benefit is just made a start. As experiments have indicated, greenmanuring and pasturing on fallow land gained higher speed of biomass accumulation, also resulted in larger capacity of infiltration than upland rice and paddy.
URI: http://hdl.handle.net/11455/27541
Appears in Collections:土壤環境科學系

文件中的檔案:

取得全文請前往華藝線上圖書館



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.