【作者】 张卫华；

【导师】 魏朝富； 谢德体；

【作者基本信息】 西南大学 ， 土壤学， 2011， 博士

【摘要】 随着世界人口的日益增长,雨养农业在为人类提供粮食和生计方面起着不可替代的作用,并且这个作用将持续下去,因此,如何提高雨水的就地利用以降低农业风险是当前的急迫任务之一。很久以前,雨水的收集利用就在世界上许多国家得到了应用实践,尤其是在干旱半干旱地区,如中国的黄土高原地区,北美的西部地区,中东,撒哈拉大沙漠的干旱地区。但是,在湿润地区这方面的研究甚少,尤其是西南丘陵山地,几乎没有系统的雨水收集利用体系。另外,目前仅有的一些研究雨水收集利用体系,主要集中在农村生活和家庭用水方面,少有农业方面的应用,比如传统的格田,非洲的利用小型坑洼的耕作方式等；几乎很少研究是利用土壤的蓄水能力,并结合当地的地形特点及特殊的水文过程来对雨水资源进行利用从而减缓季节性干旱的。在西南丘陵山地的重庆地区,尽管年降雨量较丰富,但是由于其时空分布极不均匀,季节性和区域性干旱时有发生,给当地农业带来了巨大损失。因此,分析土壤水分的特性并重点利用土壤蓄水能力对雨水进行就地利用,在土壤水分无法满足作物需求情况下利用蓄水设施蓄积多余径流进行补充灌溉,对保障农业用水以及减低季节性干旱带来的损失具有重要的科学和现实意义。通过利用实地监测的土壤数据及采集的土壤样品,分析了土壤的物理特性,尤其是那些影响雨水就地利用如下渗和蒸发特性及其影响因素、土壤水分随时间的变化特性等。然后结合GIS,基于SCS-CN的水文模型,对研究区的降雨径流过程进行了模拟,以此来获取径流收集潜力的相关信息。最后,根据不同的土地利用情况,计算作物需水,并与来水进行平衡计算,从而计算出收集径流的蓄水池尺寸来应对季节性干旱时进行补充灌溉。主要研究结果如下：(1)根据重庆近100年来的年降水量分析,以降水的距平百分比判断,2006年是干旱少雨年份,而2007年被认为是正常年份。对土壤水分的年际变异,如果按照变异系数CV<10%为弱变异,10%<CV<30%为中等变异,以及CV>30%为强变异来划分,2006年的0-10cm,10-20cm,20-40cm三个土层的土壤水分变异均为中等程度变异。而对于正常年的2007年,0-10cm土层为中等程度变异,另外两个土层为弱变异。因此,对于湿润地区,干旱年份土壤水分的年际变异较大,而湿润年份相对较小。对月际变化,土壤水分在2006年的总体趋势处于下降趋势,尽管在5月至6月份之间有稍微的恢复。这主要是因为高温、蒸发以及作物生长旺盛,降雨稀少不能有效补充土壤水分导致的。而在2007年,土壤水分变化像“W”型波动,从3月到5月,由于作物生长旺盛,土壤水分快速递减,从5月份到7月份,尽管作物生长依旧旺盛,但是随着降雨增多,土壤水分含量升高。从7月份起,水分又开始降低。总之,土壤水分的月际变化可以分为两个阶段：3月-5月作物耗水少雨阶段,以及6月到9月的波动阶段。对于旬际变化,无论是干旱年份还是正常年份,土壤水分变化都比较剧烈,但是由于2006年干旱的影响,该年的变化要远远比2007年大。对于土壤水分的概率分布,无论是干旱年份还是正常年份,都出现单峰形状,只是峰值出现的位置及峰的阔度不同年份、不同土层有所不同。(2)在土壤水分收支平衡的诸多影响因素中,砾石的出现改变了土壤物理性质如可利用土壤水分、土壤下渗、蒸发过程,而这些又是控制利用土壤水分的重要因素。本研究在实验室测定了研究区土壤的这些水力性质,不同砾石含量对三种不同紫色土的影响,尤其是砾石含量对土壤水分的补充和耗散参数,即下渗和蒸发的影响。结果表明：对于下渗率,无论是垂直下渗还是水平扩散,均是灰棕紫泥>红棕紫泥>棕紫泥；另外,随着砾石含量的增加,累计下渗量递减,而下渗时间先减后增。下渗率也随着距离变化,在0-5cm范围内,初始下渗率随着砾石含量的增加而升高,在5-10cm范围,下渗曲线的斜率随着砾石含量的增加而增大。然而,随着距离的增加,斜率逐渐递减直到最后达到稳定值。砾石的出现也降低了土壤水分含量,最小值出现在当砾石在土柱的表层时；当砾石和土壤混合时,土壤水分含量随着砾石的增加而减少。在保持恒定温度情况下,当砾石覆盖在土壤表层时,累积蒸发量和蒸发率在最小；对其他样品,累积蒸发量随着砾石含量的增加而减少,而蒸发率在前四天随着砾石含量的增加而增加。(3)本研究中利用GIS和SCS-CN耦合的水文模型来计算研究区子流域中的径流量。CN值是通过土壤类型,土地利用等信息用权重平均方法计算出来的,降雨是利用的2000-2005年的月平均降雨数据。计算结果表明：研究区径流集中在4月至10月份,这7个月的径流量占了全年径流量的90%以上,然而7月和8月的径流量只占了全年径流量的很小一部分,可能是和此间的蒸发、下渗和蒸腾有关。径流年内分配不均匀系数分别是0.97,1.25,1.15,1.10,1.24和0.97。而由六年平均月降雨计算出来的年内分配不均匀系数为0.9.明显低于任一年的值,这表明了每年的径流年内变化是不同的,2000年和2005年相对较小,而2001年和2004年相对较大。六年的平均径流系数是0.49,变差系数是0.27.较大的径流系数和南方湿润地区的产流规律吻合,然而,和重庆地区经常引用的0.51的径流系数相比,实际情况还是有所差异的,比如这六年中,最小的2001年的径流系数只有0.4,最大的2004年为0.54。因此对于要求较高的径流计算,需要比较详细的模拟降雨径流过程。对丘陵区农业而言,局地径流调控和收集是一个关键问题。坡改梯工程可以提高雨水在土壤表面的停留时间从而增加下渗量。本研究中,通过降低三个不同范围的坡度,土层厚度增加了20cm至30cm。雨水停留时间也分别增加了2秒到60秒不等。土壤蓄水量也增加了168m3/hm2至245m3/hm2,这极大程度的帮助了降雨的就地利用程度。此外,根据现有的水库和蓄水设施,结合坡改梯工程,径流收集潜力分析和作物需水量,利用水量平衡法需要增加的蓄水设施来进行补充灌溉。(4)为减轻干旱带来的影响,本研究从气象干旱和土壤干旱两个方面进行了分析,对不同程度的干旱概率进行了计算,结果表明春旱和伏旱的概率都超过了60%。对于不同土地利用方式和坡度来说,作物需水量和土地利用方式和坡度密切相关。由于土壤是作物需水的天然水库,因此对不同坡度、坡位的土壤水分特性和土壤水库容量进行了分析。通过对供水(如降雨、干旱特性、土壤水库)、需水(作物蒸散发)的分析,进行了水量平衡的研究。研究结果表明,对于荣昌县大坝村的旱地农业来讲,需要27个蓄水池来规避季节性干旱,其中8个容积为100立方米,19个为50立方米。对于稻田,需要修建5.05hm2的囤水田来蓄积1.57*104m3水量来进行补充灌溉。这些囤水田的布局应根据稻田分布情况及地形情况分散布局,以方便灌溉。总之,雨水的就地利用是减轻季节性干旱的有效途径,可以提高农业用水的可靠性。降雨和径流的收集利用在干旱半干旱地区应用甚广,然而,近年来,在湿润地区也开始应用,即使是在有较好的灌排体系的地区。在重庆丘陵山地,季节性干旱频繁发生,并且农业供水设施不足。因此,对土壤水分动态及其影响因素的了解可以很好的帮助雨水的就地利用。降雨径流过程可以帮助计算径流收集潜力,坡改梯可以提高土壤蓄水能力和增加雨水下渗,从而提高雨水利用效率。除了下渗和蒸发外的降雨径流可以被收集于小型蓄水设施里用来短期干旱时的补充灌溉,这样就可以降低雨养农业由于缺水而导致的风险。更多还原

【Abstract】 Rainfed agriculture plays and will continue to play a dominant role in providing food and livelihoods in the world with an ever increasing population. How to improve in-situ rainwater use and hence reduce the agricultural risk is an important task nowadays. Rainwater harvesting has been practiced for many years in many areas of the world, especially, in arid and semi-arid regions like the Loess Plateau, North America, Middle-East, sub-Saharan Africa etc. However, there is little research done in humid area, even less done in the hilly area in Southwest China. Furthermore, the existing research is mainly focus on rural domestic use in north part of China, also some in agricultural but only use traditional practices like contour bunds, pitting cultivation etc. No research was done by using the soil water storage characteristics, combined by typical hilly area’s hydrological process to mitigate the seasonal drought. In the hilly area of Chongqing, the annual rainfall is relatively rich. However, with its spatial and temporal variation, the regional and seasonal drought occurs frequently, which leads to serious agricultural losses every year. Hence, the core scientific aspect in this research is in-situ rainfall and runoff harvesting through analyzing the soil water characteristics, integrating the GIS and hydrological model, to supplement rainfed agriculture during dry spells in hilly area of Chongqing. The study has a profound scientific and practical meaning in mitigating the seasonal drought in this region.By using the monitored soil data and measured soil samples, the soil characteristics, especially those closely related with in-situ rainfall use like temporal variation of soil water content, infiltration and evaporation, were analyzed. Then by using GIS, SCS-CN based hydrological model, the rainfall runoff process was simulated to predict runoff potential for harvesting. Finally, under different land use pattern, the crop water requirement was calculated and size of storage ponds were calculated for supplementary irrigation based on water balance approach. The main research findings are as follows: 1) According to the recently 100 years rainfall data, taking the precipitation anomaly percentage as the criteira, the year 2006 was a dry year and 2007 a normal year. For the dry year, the variations of soil moisture in all three layers (0-10cm,10-20cm, 20-40cm) were medium (10%<CV<30%); while for the normal year, the variation in the layer of 0-10cm was strong (CV>30%), and weak in the other two layers (CV<10%). Hence, in the humid area, the seasonal variation is large for dry year and relative small for wet year. For monthly variation, in the year 2006, the overall trend of soil moisture was declined, although there was a slight recovery from May to June. This was mainly due to the high temperature and evaporation, the actively crop growth worsen the situation. In the year 2007, the soil moisture fluctuates like a "W", from March to May, high crop water requirement made soil water decreased rapidly, from May to July, although crop water requirement was still high, with the increasing rainfall, the soil water content increased. From July, similar situation occur. In summary, monthly variation can be divided into high crop consumption and low rainfall period (March to May) and fluctuation period (June to September). For decadal variation, no matter for dry year or normal year, there was a drastic fluctuation of the soil water content. However, due to the impact of 2006 drought, the fluctuation of soil moisture in 2006 was much bigger than that in 2007. For five-day variation, the upper layer (0-10cm) was the most sensitive layer, and there were clear difference between different layers.For the probability distribution of soil moisture, no matter for the dry year or normal year, all the distribution in three layers have a single peak shape. However, the location of the peak appeared at different position in different layers and years, similar with the band of the peak.2) Among the various factors, the presence of rock fragments modifies the soil physical properties such as available water content, infiltration and runoff susceptibility. The results of laboratory experiments showed the effects of rock fragments contained in three different purple soils of the Sichuan basin of southwest China. The experiments investigated how these rock fragments alter the soil’s physical, chemical and agronomical characteristics such as infiltration and evaporation. We found that the infiltration rate, whether horizontal or vertical, in the three soils has the following order: grey brown purple soil>reddish brown purple soil>brown purple soil. With increasing rock fragment contents the accumulated infiltration decreases, while the total time decreases first and then increases. The minimum occurs at approximately 10% to 20% of fragment content by weight. The infiltration rate also changes with the distance. In the 0-5cm range, the initial infiltration rate increases with increasing rock fragment contents, in the 5～10cm range, the slope of infiltration curve increases with increasing rock fragment contents. With increasing distance, the slope gradually decreases and finally reaches a stable value. The presence of rock fragments reduces soil water content, with the minimal value appearing when the rock fragments where on top of the soil column, while decreasing with increasing rock fragments with soil samples mixed with fragments. Under the constant 40℃temperature, the accumulated evaporation and evaporation rate are minimal for soils covered by rock fragments, and the accumulated evaporation decreases with increasing rock fragment for other soil samples. However, the evaporation rate increases with increasing rock fragments in the first 4 days and decreases thereafter.3) The GIS coupled with SCS-CN model was applied in this study to calculate runoff in each micro-catchment. The monthly average rainfall from 2000～2005 were analyzed and used to calculate the runoff. The CN number of each catchment is calculated based on its soil groups, land use areas using weighted mean method. The results showed that with different CN number, the runoff is different even with the same soil conditions. The runoff concentrated in the period of April to October which stands more than 90 percent of total annual runoff. However, the runoff from July and August only take a small part of total yearly runoff, which is closely related with the average rainfall of these two months; possibly there is certain relation with evaporation, transpiration, and infiltration. The coefficients of uneven distribution for annual runoff in the year of 2000 to 2005 were 0.97,1.25,1.15,1.10,1.24 and 0.97 respectively, however, the coefficient got from 6 years monthly average data is 0.90, which is much smaller than any of the value in the years. The results illustrates that the variation of runoff is different in the years, with 2000 and 2005 being relative small, and 2001 and 2004 rather big. The average runoff coefficient during this 6 years’period is 0.49, with the coefficient of variation of 0.27. The high runoff coefficient is consistent with the fact of high runoff in southern part of China. However, compared with the 0.51 runoff coefficient generally used in Chongqing area, there is in fact some difference. For example the smallest value is 0.4 in 2001, and a large value of 0.54 in 2004. Hence, in the study for more accurate runoff calculation, the rainfall runoff process should be further studied.It is a key issue for regulating and harvesting local runoff in hilly area. The slope to terrace project can increase the residence time of rain water on the soil surface and hence increase the infiltration time. In this study, three categories of slope had been lowered. The soil layer had been increased 20-30cm, and the residence time increased 2 to 60 seconds. The soil water storage increased 168 to 245 m3/hm2 for different slope change, which greatly helps the crop to abstract water from soils.In addition, according to the existing reservoir and water storage facilities, the slope to terrace project to increase soil water storage, and runoff potential analysis, crop water requirement, the water balance was made to design the size of water storage cisterns to supplement irrigation in each sub-catchment. The size of the cisterns is usually 100m3, with some small ones in catchments which need less water for crop requirements.4) In order to mitigate the seasonal drought, in this study we analyzed the drought characteristics from both meteorological and soil aspects. The probabilities of drought with different severities were calculated and compared with the stationary distribution probabilities. The results showed that both the spring drought and dog-day drought had the probability greater than 60%. For crop water requirements under different land use pattern and slope, we found that there was close relationship between these variables. As soil is a natural reservoir to store water for plant use, we also investigated the soil water characteristics, different soil reservoir capacities and their variation at different locations on the slope. After the analysis of the supply side (rainfall, drought characteristics and soil reservoir) and demand side (crop evapotranspiration) were performed, a balance was tried between supply and demand. We found that in the dryland agriculture, 27 storage cisterns was needed to mitigate seasonal drought in the study area; among which 8 should have a capacity of 100m3 and 19 with a capacity of 50m3. For paddy field, we need to establish bunds to form a totally 5.05hm2 fields to store 1.57*104m3 water. However, the fields should be sparsely distributed in order to easily supply water to different farms. Implications are that detailed meteorological and hydrological data are needed to make the results sound. However, for a local area, these data are usually difficult to obtain. Assumptions were made for some special situations in this study.In summary, in-situ rainfall use is an efficient way to mitigate seasonal drought and hence improve reliability of water to agriculture. Rainfall and runoff harvesting is commonly used in arid and semi-arid regions, however, interest in rainwater harvesting increases in humid regions, even areas containing well developed water supply infrastructures. In the hilly areas of Chongqing, seasonal droughts occur frequently and the water supply infrastructure is quite poor. Therefore, a better understanding of soil water dynamics, controlling factors influencing soil water recharge and discharge is critical for better using rainwater in-situ. Rainfall runoff process can help identify the potential of harvesting. Slope to terrace project can improve soil water storage and hence increase rainwater use efficiency. The remaining rainwater after infiltration and evaporation can be stored in small-scale reservoirs or cisterns for supplementary irrigation during seasonal drought, which reduces the risks of rainfed agriculture apparently in this region.更多还原