【作者】 熊友胜；

【导师】 魏朝富； 谢德体； 何丙辉；

【作者基本信息】 西南大学 ， 土壤学， 2013， 博士

【摘要】 三峡库区农业以旱作粮食作物为主,坡耕地是库区主要的农业生产用地,降水是该区域农业生产的主要来源,降水总量大,但降水时间分布不均,径流损失大,该区域季节性干旱十分突出,基本上是十年九旱,干旱缺水严重制约该区域农业的持续发展。紫色土坡耕地是库区重要的耕地资源,耕作频繁,人为扰动剧烈,水土流失严重,以雨养农业为主,土层薄,蓄水保肥能力差,水分亏缺是作物生长的主要障碍因子。如何合理调节利用降水资源,解决该区域季节性干旱与植物生长发育协调性差的矛盾,作物耗水特性和农田水量平衡研究就显得特别重要。国内外有关农田水量平衡的研究早已不只局限于理论本身,而是将相关理论用于指导农业生产实践,并取得了丰硕的成果,但是,国内的研究主要集中于干旱、半干旱的华北平原和黄土高原地区,而在南方季节性干旱地区,特别是三峡库区紫色土坡耕地水量平衡的研究未见详细报道。目前,紫色土坡耕地水量平衡要素研究中径流的研究较多、较深入,但在地表径流模拟计算方面仍过程复杂,同时对最主要也是最难确定的农田蒸散量这一平衡要素,特别是涉及农田蒸散量的确定方面未见详细研究。为此,本研究针对紫色土坡耕地水量平衡研究中的不足,选择重庆市开县竹溪镇石碗小流域坡耕地(坡式梯地)为研究对象,以紫色砂岩、泥岩发育形成的中性紫色土为供试土壤,开展坡耕地水量平衡分析研究,为今后全面开展三峡库区紫色土坡耕地水量平衡的监测和预测等深入研究打下基础,同时希望能为库区坡耕地分类改造,坡耕地集、蓄、供、管的降水资源化集成利用体系的建设提供科学依据。试验研究以为三峡库区季节性干旱问题的解决提供可靠数据为出发点,以2007-2012年盆栽试验、测坑试验和径流小区试验观测数据为依据,以坡面产流汇流理论、水量平衡原理为理论基础,采用试验观测、理论分析计算等方法,对三峡库区紫色土坡耕地水量平衡要素进行了探讨,对常规种植模式下单作小麦、油菜、玉米、红薯和小麦/玉米‖红薯、油菜/玉米‖红薯间套作模式的作物耗水特性和坡耕地水量平衡状况进行初步研究。主要研究结果如下(1)水分下渗过程主要受控于降水强度的变化过程2007-2012年作物生长期内降水呈现出夏秋多、冬春少,夏秋季多暴雨,降水量集中的特点。6-8月古年均降水量的43.4%,7、8月暴雨量占月均降水量的55.2%、44.5%。降水分布特征参数(σ、Cv、Cs)分析结果表明作物生长期内降水在时间分布上呈不均匀性,One-Sample K-S检验表明作物生长期年内各月降水旱正态分布。坡耕地单点土壤水分下渗在时空分布上存在变异性,坡面不同位置初始下渗率和稳定下渗率具有较大差异,下渗历时60min拟合第1分钟末下渗率8.03～20.83mm.min-1(Kostiakov公式拟合值),稳定下渗率0.76～3.81mm.min-1,均值1.42mm.min-1(实测值),耕作对土壤下渗影响显著。初始下渗的0～10min时段,土壤下渗率迅速下降,而且下降幅度较大,10～25min时段,下降幅度明显减小,下渗历时25min后,逐渐趋于稳定,在27.7～35.4min内达到稳定下渗阶段。单点下渗特性可用Kostiakov公式来描述,公式中第1分钟末的下渗速率与初始含水量呈高度负相关关系(可用指数函数关系式表达),下渗率和下渗量表达式如下：自然降水条件下土壤下渗特性与降水强度有密切的关系,下渗率的变化过程主要受控于降水强度的变化过程,坡面下渗特性受诸多因素的影响,在空间上和时间上都呈现出不稳定和不连续性变化。暴雨时段前后降水强度不能满足下渗率要求,降水全部下渗,即f(t)=i(t),降水强度大时,土壤的下渗率大,降水强度小时,下渗率则小。在暴雨时段内,前时段降水强度大时,即i(t)≥fp(t),土壤的下渗率大,即f(t)=fp(t),但维持较高下渗率的时间较短,以后就逐渐减小；在暴雨时段内,前时段降水强度小时,即i(t)≤fp(t),降水全部渗入土壤,即f(t)=i(t),此时下渗率小,当降水强度增大到i(t)≥fp(t)时,下渗率达最大f(t)=fp(t),随后又逐渐减小,维持这种较低或较高下渗率的时间长短,取决于暴雨时段内前时段小强度降水历时。(2)分时段净雨模拟方法能较好反映的坡面产流过程不同降水特性、雨前土壤含水量和耕作状况等对产流时间、产流量和径流成分组成有影响。雨峰靠前、强度大的降水,地表径流出流时间早而且地表径流量大,若土壤初始含水量高,暴雨时段后小强度降水历时长,则壤中流产流量也较大；雨峰靠中的降水,暴雨前土壤含水量高,即使是中等强度的降水,如果降水量大、历时长也将产生大量的地表径流和壤中流；雨峰靠后的降水,如果暴雨时段内前期降水未使土壤含水量明显提高,降水首先满足表层土壤的缺水量,当降水强度超过下渗能力时开始产生地表径流,若降水量不大、暴雨历时较短,降水损失量大,则地表径流和壤中流出流量偏小。地表径流主要为暴雨产生的超渗产流,7、8月以地表径流为主,月均径流系数分别为0.36、0.30;5、6月和9月以壤中流为主,耕作后土壤以壤中流产流为主,试验区坡耕地壤中流较发育,壤中流径流系数0.15～0.34。壤中流的产生与降水强度无直接关系,与土壤前期含水量、土壤中是否存在自由重力水和暴雨前后小强度降水特性有关。2008～2012年降水径流主要出现在5～9月,月均径流系数分别为0.15、0.28、0.36、0.30、0.07,5～9月月均降水径流关系可表示为：R=0.559P-44.03R2=0.948以坡面水流运动波理论、流量过程的倍比假定和叠加假定原理为理论依据,以暴雨时段平均下渗率代替瞬时下渗率的分时段净雨模拟方法能较好反映自然降水条件下的地表径流产流过程,简化了模拟过程,可操作性强,模拟结果较好,模拟地表径流量与实测值相吻合。雨峰靠前、靠中和靠后的三次降水模拟产流过程与实际观测结果大致吻合,模拟地表径流量分别为34.4mm、39.7mm和10.2mm,与实际观测结果也大致相同。(3)农田蒸散量采用作物系数法确定太阳辐射采用Hargreaves公式计算。标准的Penman-Monteith计算公式以能量平衡和水汽扩散理论为基础,既考虑了作物的生理特征,又考虑了空气动力学参数的变化,较为全面地考虑了影响蒸散发的各种因素,具有较充分的理论依据。试验区参考作物蒸散量主要由空气动力学项ET0(aero)贡献,约占70%,辐射项ET0(rad)约占30%。2001-2012年日、月、年参考作物蒸散量的变化趋势均与温度的变化趋势相同,而年参考作物蒸散量的变化趋势与日照时数变化趋势相反。相关性分析和敏感性分析结果,最高温度和日照时数与参考作物蒸散量的相关性一致,同时日照时数与最高温度具有较大的相关性,饱和水汽压差与参考作物蒸散量具有高度相关性,而饱和水汽压差是温度和相对湿度对参考作物蒸散量影响的综合反映；最高温度的变化对参考作物蒸散量的影响最大,敏感性最强,而日照时数的变化对参考作物蒸散量儿乎无影响,敏感性最弱,结果体现出试验区温度对参考作物蒸散量的影响远大于日照时数,而日照时数对参考作物蒸散量的影响主要通过温度间接体现；最高温度(Tmax)、最高相对湿度(RHmax)和最低相对湿度(RHmin)是参考作物蒸散量三个最主要的敏感性因子。显著性分析结果,Angstrom公式、Hargreaves辐射计算公式和重庆地区拟合经验公式计算太阳辐射所得日、月、年参考作物蒸散量值均无显著差异(α=0.05)。确定采用Penman-Monteith公式计算参考作物蒸散量时,采用Hargreaves公式计算太阳辐射。作物系数采用双值作物系数。单值作物系数计算简单,各个阶段取其平均值,反映不出作物蒸腾量和土壤蒸发量的大小：双值作物系数不仅能反映各生长阶段作物蒸腾量和土壤蒸发量的大小也能反映基本作物系数(Kcb)和土壤蒸发系数(Ke)的时程变化特征,而且双值作物系数的日变化过程,有利于分时段(如：日、旬、月)作物系数的确定以及间套作模式作物系数的确定,从而有利于坡耕地各旬、各月水量平衡分析。经调整后小麦、油菜、玉米、红薯全生长期作物系数分别为0.92、0.93、1.06、0.98。权重系数法确定问套作模式作物系数。间套作模式田间结构为一人工复合群体,有主、副作物之分,涉及不同作物的不同生育期,单位面积上不同作物在田间结构中占的面积比例将直接决定其农田蒸散量的多少,而权重系数能反映出单位面积上不同作物在田间结构中占有的面积比例。采用权重系数计算小麦/玉米‖红薯、油菜/玉米‖红薯间套作模式下各月作物系数,从而解决了利用作物系数法计算农田蒸散量时,间套作模式下作物系数难确定的问题。具体为,以作物幅宽与间距之和占带宽的比作为权重系数,分别乘以间套作模式的不同作物的作物系数,取其和作为间套作模式的作物系数,即：式中：K1为间套作模式的作物系数：Kci为间套作时第i作物的作物系数;fi为第i作物的权重系数；li为第i作物的幅宽；d、L—间距、带宽；i为间、套作时作物种类(i=1、2……n)。(4)作物耗水主要受土壤水分胁迫降水量(P)是坡耕地土壤水分的主要来源,农田蒸散量(ETc)是坡耕地土壤水分主要损失项,径流量(R)、渗漏损失量(F’d)主要产生于5-9月,试验区坡耕地水量平衡简化模型可表达为：P-ETc-R-Fd’=ΔW。标准状态下,坡耕地水量平衡值为负值(⊿W<0),水分均处于亏缺状况。2007-2012年小麦、油菜、玉米和红薯单作以及小麦/玉米‖红薯、油菜/玉米‖红薯间套作全生长期作物需水量均值为359.4mm、359.1mm、684.4mm、841.3mm、1178.3mm、1158.7mm,亏缺量为136.2mm、135.9mm、298.7mm、336.4mm、451.9mm、432.3mm,作物全生长期水量平衡值为负值(⊿W<0),十壤水分处于亏缺状态,水分亏缺时期主要为2-4月和7-9月,与试验区春旱、夏伏旱和秋旱旱情发生时期相一致。自然降水条件下,坡耕地作物耗水量受土壤水分胁迫的制约。在水分胁迫下,2008-2012年小麦、油菜、玉米和红薯单作全生长期作物耗水量均值为255.6mm、256.0mm、553.3mm、621.8mm,全生长期水分胁迫系数(Ks)均值分别为0.85、0.84、0.87、0.82。作物不同生长阶段受水分胁迫的影响及程度不一,生育中期、生长后期受水分胁迫严重。作物发生水分胁迫现象受气象因子变化的影响显著,发生水分胁迫时的土壤含水量有随参考作物蒸散量的增大而增高的趋势,即当参考作物蒸散量较大时发生水分胁迫的土壤相对含水量越高,当参考作物蒸散量较小时发生水分胁迫的相对含水量较低。小麦、油菜发生水分胁迫时土壤相对含水量约60%左右；玉米、红薯的相对含水量范围变幅大,玉米在61.1%-76.8%范围内均可发生水分胁迫现象,对应参考作物蒸散量范围3.68-10.76mm.d-1,红薯为51.2%-80.3%,对应参考作物蒸散量范围1.66--11.42mm.d-1。综上所述,基于坡面汇流理论、流量过程的倍比假定和叠加假定原理为基础,以暴雨时段平均’下渗率代替瞬时下渗率的分时段净雨模拟方法,能较好模拟自然降水条件下的地表径流产流过程,简化了模拟过程,可操作性强,便于实际应用；在详细分析研究基础上,确定Penman-Monteith公式计算参考作物蒸散量时,采用Hargreaves公式计算太阳辐射。确定作物系数采用双值作物系数法计算。权重系数法解决了间套作模式下农田蒸散量计算时作物系数难确定的问题；坡耕地土壤水分亏缺严重,作物耗水量主要受土壤水分胁迫。这些研究成果不仅丰富了紫色土丘陵区作物耗水量的研究内容,也为紫色土丘陵区深入开展农田水量平衡研究打下基础。但在自然条件下降水下渗特性、坡面曼宁糙率系数、短时段水量平衡以及水分胁迫条件下农田水量平衡等研究方面还有待于进一步完善。更多还原

【Abstract】 Three Gorges Reservoir Basin (TGRB) is a fragile Eco-economy area in China, where the agriculture is mainly dry farming cereals. The slope farmland is the main agriculture land use, and precipitation is the key resources for agriculture. The total precipitation is plenty; however, the uneven distribution in time and space makes the seasonal drought a serious problem in this region, which greatly hinders the sustainable development of agriculture. The slope land of purple soil, with frequent cultivation and human disturbance, is facing serious erosion. The rain-fed agriculture in this thin soil layer makes the water the key factor of productivity. Hence, how to manage the precipitation resources efficiently and match the crop water requirement with rainfall is a critical point in dealing with the seasonal drought and plant growth. Literatures shows that the farmland water balance is not only theoretically, but also practically to direct the real agricultural production. However, the research in China was mainly focused on arid and semi-arid areas like North China Plain and loess plateau, seldom seen in the southern part with seasonal drought, especially, in the TGRB’s purple soil area. Currently, the research in the slope farmland of purple soil is mainly on runoff, but less attention paid on evapotranspiration (ET), especially crop ET. In order to fill the research gap on farmland water balance in Southern part of China, this paper chose the ShiWan catchment’s slope land in KaiXian County of Chongqing as the study area, using the purple soil as the testing samples, to research on farmland water balance. This research will help the future work on purple soil water balance, its monitoring and forecasting, at the same time, hopefully it can help the categorization of slope farmland, provide the scientific background to integrating rainfall collecting, storage, supply and management in southern China part.Aiming to providing scientific basis for solving the seasonal drought problem, this study performed the pot experiment during2007～2012, pit-test experiment, and runoff plot experiment, integrating water balance, slope runoff generation theory, using field observation and theoretical analysis and calculation, the water balance and various components were analyzed in the purple soil farmland in TGRB. In order to reflect the real farmland situation, the regular wheat, oil-seed, maize and sweet potato and rotation combination of wheat/maize‖sweet potato, oil-seed/maize‖sweet potato were arranged during the experiment. The main results are as follows:(1) The infiltration process was dominated by the variation of rainfall intensity.During the growing season of2007to2012, it was shown that the summer and autumn had more rainfall, while less in winter and spring. Furthermore, there are usually rainstorms in summer and autumn and the rain were prone to cluster. The rainfall during June to August accounted for43.4%of the annual rainfall, and July and August rainfall accounted for55.2%and44.5%of the monthly average values. The statistical parameters Cv and Cs showed that the rainfall distribution were uneven during the growing season, one-sample K-S test showing that the monthly rainfall being with normal distribution.The single point infiltration varied in space and time in the slope land. The initial and stable infiltration rates had huge difference in different location of the slope land. For the60mins duration, the simulated rate in the end of1min ranges from8.03to20.83mm.min-1(Kostiakov simulation), and the stable rate ranges from0.76to3.81mm.min-1, an average value of1.42mm.min-1, and the cultivation had an important effect on infiltration. The infiltration rate decreased with time, during the period of0to10mins, the infiltration decreased fast with large amplitude; for the period10to25mins, the decreasing rate slowed down. After25mins, the rate starts to be stable and reached constant during27.7-35.4mins. The infiltration process can be simulated by Kostiakov formula, the infiltration rate a in the end of1min inversely correlated with initial soil water content, the infiltration rate and accumulated infiltration can be simulated by:The soil infiltration characteristics had a close relationship with rainfall intensity, the infiltration rate was dominated by rainfall intensity, and furthermore, the slope land infiltration was affected by many factors, which showed an unstable and non-continuous situation on both temporal and spatial scales. The rainfall cannot meet the infiltration during the beginning and end of the storm, which means f(t)=i(t) during this period. During the storm, when i(t)≥fp(t),f(t)=fp(t), but with short time interval and decrease with time. When i(t)≤fp(t), f(t)=i(t), all the rain infiltrated. With rainfall increase, the infiltration increases until to the maximum, and then decreases again. This process was dependent on the rainfall duration, infiltration characteristics and rainfall intensity.(2) Simulation with different time-interval can best reflect the runoff generation process during natural rainfall condition.The various factor like rainfall characteristics, antecedent soil water content and cultivation method had impact on the amount, time and component of runoff. With early peak and high intensity rainfall, runoff came early with large amount; if with high initial soil water content and long duration of low intensity rainfall after storm, the interflow accounted a significant amount. With the peak rainfall in the middle, high initial soil water content, even with normal intensity rain, if it lasts long enough, there would also be high runoff and interflow. However, with late peak rainstorm, if the previous rainfall being not replenishing the soil clearly, the rain will first satisfy the soil water moisture. Under this situation, only when the rainfall intensity exceeds the infiltration rate, there occurs runoff. If with low rainfall and short time, the loss will be huge and runoff and interflow are less.The surface runoff was mainly from infiltration excess runoff, and the runoff was during July and August. There are quite developed interflow in the study area, which can be found in May, June and September with a runoff coefficient of0.15-0.34. This interflow had no direct relation with rainfall intensity; however, it is related with antecedent soil moisture, saturation soil water and low intensity rainfall. The runoff mainly came during May to September and the monthly runoff coefficient is0.15,0.28,0.36,0.30and0.07separately. The relationship between rainfall and runoff can be expressed as: R=0.559P-44.03R2=0.948Based on the kinematic wave theory, proportionality and superposition theory, the time-interval runoff simulation can best reflect the runoff generation process. The method can clarify the initial and end time point, at the same time simplified the whole process. The results showed that there was perfect match between simulated values with measured data. The simulated runoff process matched well for the precipitation peak at the front, middle and rear of the rainfall event, the simulated runoff amount being34.4mm,39.7mm and10.2mm, which was in accordance with the observed data.(3) The evapotranspiration (ETc) was calculated by crop coefficient.The radiation was determined by Hargreaves formula. The standard Penman-Monteith formula is based on energy balance and water-vapor diffusion process, which not only accounts for physiologic characteristics, but also integrates aero-dynamics variations. The formula takes into consideration of various factors influencing ET, and has strong theoretical background. The reference evapotranspiration (ETo) have two parts:one is radiation part ET0(rad), another is aero-dynamic part ET0(aero).The (ET0) was mainly contributed by ETo(aero), accounting for70%, while with ET0(rad) being30%. During the period of2001～2012, the monthly and annual ET0(aero), ET0varied with a similar trend with temperature, while opposite with sunshine hours. Correlation and sensitivity analysis showed that:the climatic factors (temperature and sunshine hours) displayed significant correlation with ETo, same correlation being found between sunshine hour and maximum temperature; while the saturation vapor pressure deficit was the comprehensive reflection of crop reference evapotranspiration from temperature and relative humidity. The maximum temperature (Tmax) is the most sensitive factors for ET0while the sunshine hour the least sensitive. The results showed that the effect of temperature had far exceeded the effect of sunshine hour, which was only an indirect factor. Among the factors, the maximum temperature (Tmax), maximum relative humidity (RHmax), minimum relative humidity (RHmin) are three most sensitive factors for ET0. The significance analysis showed that:Angstrom, Hargreaves formula and Chongqing empirical formula had significant difference with the level of α=0.05in calculating daily ET0(rad) and annual ET0(rad), while monthly ET0(rad) had no significant difference. While the radiation (Rs) method to calculate the three methods daily ET0, monthly ET0and annual ETo had no significant different with α=0.05. According comprehensive analysis, the reference evapotranspiration was determined by Penman-Monteith method, the radiation by Hargreaves formula.The determination of crop coefficient was done by dual-crop coefficient. Single crop coefficient is simple which takes the average of the three growing period of a crop. However, the results from different hydrological years have clear difference due to the initial growing period coefficient greatly affected by rainfall (or irrigation) situation. Dual crop water coefficient can not only reflect the transpiration and soil water evaporation, but also the temporal variation of the basal crop coefficient (Kcb) and the soil evaporation coefficient (Ke). Furthermore, the daily variation of dual crop water coefficient will help the determination of various time intervals (e.g. day’s, decade’s or month’s) crop coefficient and the coefficient during rotation and co-existence for two or more crops. By adopting the dual crop coefficient, the crop coefficient of wheat, oil seed, maize and sweet potato was0.92,0.93,1.06, and0.98.The determination of intercropping patterns crop coefficient was done by weighted sum method. For intercropping plantation, it is more complex as there is main crop and associate crop, furthermore, the two or more crops are not planted simultaneously, which makes it difficult to determine the coefficient. Here, a weighted sum method was proposed to calculate the monthly crop coefficient of wheat/maize‖sweet potato, oil seed/maize‖sweet potato, and hence solve the problem of difficulty to determine the coefficient issues. More specifically, based on the range and space of crops as the weight, multiplied by the crop’s coefficient, the sum will be used as the total crop coefficient, that is:Where K’being the coefficient of intercropping crops; Kci being the ith crop’s coefficient;fi being the ith crop’s weight;li being ith crop’s range; d, L being space and range; i being the crop species (i=1、2......n).(4) The crop water consumption in the purple slope farmland limited by soil water stress.Precipitation (P) is the main water source and evapotranspiration (ETc) is the main water consumption item in the purple slope farmland. The runoff (R) and deep percolation (Fd’) was during May to September. According the theoretical analysis, the water balance in the purple soil farmland is: P-ETc-R-Fd’=ΔWsUnder the standard situation, the water budget was negative(ΔW<0), there was water shortage. the crop water requirement for wheat, oil seed, maize, and sweet potato, and the intercropping of wheat/maize‖sweet potato, oil seed/maize‖sweet potato was359.4mm,359.1mm,684.4mm,841.3mm,1178.3mm, and1158.7mm separately, with a shortage of136.2mm、135.9mm、298.7mm、336.4mm、451.9mm、432.3mm, for each of them. The water budget was negative (ΔW<0), there was water shortage for the whole growing period, the main water shortage period was during February to April and July to September, coinciding with the spring drought, summer drought, midsummer drought and autumn drought situation in the region.Under natural precipitation situation, the crop water consumption limited by soil water stress. Under water stress situation, during2008to2012, the crop water requirement for wheat, oil seed, maize, and sweet potato was255.6mm、256.0mm、553.3mm、621.8mm, separately, with a water stress coefficient of0.85,0.84,0.87,0.82each. The degree of water stress in different growing season were different, there was serious water stress during middle season and late season. During water stress, the soil water content increased with increasing ET0, which means that with high ET0, the critical soil water content is high during water stress and vice versa. For wheat and oil seed, when the soil water content reached60%of field capacity, water stress occurred. For maize and sweet potato, the critical soil water content was most clearly affected by meteorological factors and the range of soil moisture being large as61.1%(ET0=3.68mm.d-1)～76.8%(ET0=10.76mm.d-1) for maize and51.2%(ET0=.66mm.d-1)～80.3%(ET0=11.42mm.d-1) for sweet potato.In summary, based on the kinematic wave theory, proportionality and superposition theory, and by using the average infiltration rate for a rainfall interval instead of instant infiltrate rate, the time-interval runoff simulation can best reflect the runoff generation process. The theory is not only simplified the simulation process, but also being easily applicable for real world situation. After further research, the Penman-Monteith formula was suggested to calculate ETo and Hargreaves formula to calculate radiation, and the crop coefficient was calculated by dual crop coefficient method, and weighted sum coefficient solved the problem of dual or rotation mode of crops to fulfill the research gap in purple hilly area. The soil water deficit in the purple slope farmland was extremely serious, and the crop water consumption limited by soil water stress. The research results not only enrich the water consumption issue, but also laid a strong base on farmland water balance study. However, the infiltration characteristic under natural precipitation, slope land Manning coefficient, water balance in short time interval and water balance under water shortage should be further explored.更多还原