【作者】 高云飞；

【导师】 赵传燕；

【作者基本信息】 兰州大学 ， 生态学， 2016， 硕士

【摘要】 草地占世界陆地面积的25%左右,是生态系统中重要的组成部分,而祁连山自然保护区中草地面积占总面积的比例达到了50.54%,亚高山草地分布最广,是草地生态系统的主体,在祁连山区生态系统服务功能中发挥着重要作用,在区域水循环中担当重要角色。本文以黑河上游天老池流域作为研究区,该流域位于祁连山区深处,降雨较多,植被较为丰富,具有涵养水源的重要作用。亚高山草地占该流域总面积的7.45%,过度的放牧对亚高山草地生长影响较大,从而影响草地植被的蒸散发,所以对亚高山草地蒸散发的研究对当地放牧具有重要的指导意义,也为当地有关部门提供一定的参考。本研究利用自制小型蒸渗仪(lysimeter)对研究区亚高山草地蒸散发进行观测,主要包括:1、每天8时和20时进行日尺度观测(包括蒸发皿水面蒸发观测);2、每周的同一天(如果由于天气原因,比如降雨,后推一天)进行小时尺度观测(包括蒸发皿水面蒸发观测);3、蒸散发与气象因子的关系;4、蒸散仪观测值和涡度数据计算蒸散发特征及其与模型模拟值的关系;5、用改进作物系数计算方法计算研究区亚高山草地作物系数,并用观测值进行验证。主要得出以下结论:1.在月尺度上,从4年的平均看,6月最大,其次是8月,7月最小;6、7和8月的日均蒸散发量都是在2013年最大,其次为2011年,然后是2014年,2012年最小;2011和2012年中,每月的日均蒸散发量大小顺序为6月>8月>7月;在2013年中,8月份日均蒸散发量最大,其次为6月份,7月份的日均蒸散发量最小;但在2014年中,7月份的日均蒸散发量最大,8月份和6月份的相对较小。草地日均蒸散发量成波动状态,与降雨有一定的关系,降雨在一天内出现的时间不同,对蒸散发的影响也不尽相同。日蒸散发量与环境因子之间的相关关系为:风速>饱和水汽压差>相对湿度>净辐射,R2分别为0.4174、0.3598、0.3461和0.3322,四个气象因子对日蒸散发量都有较大的影响。2013年9月16日到2014年9月6日涡度数据计算得到的蒸散发量,总体上呈波动趋势,非生长季(十月中旬到次年四月中旬)蒸散发量较小,基本在1 mm/d以下,平均日蒸散发量为0.64 mm/d,生长季(四月中旬到十月中旬)大概开始于4月中旬,蒸散发量开始有所增加,6、7和8月份基本达到最大值,并成稳定的波动状态,蒸散发量从9月开始减小,直到10月中旬,进入非生长季,生长季日均蒸散发量为2.98mm/day。非生长季降水较少,蒸散发量比较小,而生长季降水较多,蒸散发量较大,说明蒸散发量与降雨存在一定的正相关关系。2.蒸散仪观测数据下,在日尺度上,penman、fao-penman、fao-penman-monteith、priestley-taylor和fao-radiation五种蒸散发模型中,fao-penman-monteith模型的模拟效果最好,其他三种模型的模拟效果也比较好,fao-radiation模拟效果最差;在小时尺度上,priestley-taylor模型的模拟效果最好,其他模型都不适合该地区草地小时尺度的潜在蒸散发模拟。模型输入参数对蒸散发模拟效果影响强弱顺序为:辐射>气压>相对湿度>气温>风速;当地表辐射在±10%变动时,模拟值变化幅度达到20%左右;在其他参数±10%变动时,模拟值变化幅度在8%以下,说明除了辐射以外的其他参数是模型模拟较不敏感因素。3.涡度观测数据下,生长季penman、fao-penman、fao-penman-monteith、priestley-taylor和fao-radition五种模型模拟值与涡度计算值获得的决定系数相近,都在0.55左右,而五种模型产生的相对均方跟误差分别为2.0838、2.4406、1.7078、2.7240和9.7049,得出fao-radition模拟值比涡度值偏大,不适合该地区草地生长季蒸散发量的模拟,其他四个模型模拟值与涡度值稍大,可做适当调整后对该地区草地生长季蒸散发量进行模拟,调整后的四种模型产生的相对均方跟误差分别为1.2761、1.2921、1.2608和1.2684,并且散点都落在直线y=x的附近,调整后的模拟值与涡度计算值相近,调整后的四个模型与涡度数据结合,可以作为研究区生长季长时间序列蒸散发观测工具;由于研究区气候条件的原因,非生长季的蒸散发量很难观测到,涡度数据使这一困难得以解决。penman、fao-penman、fao-penman-monteith、priestley-taylor和fao-radition五种模型模拟值与涡度计算值获得的决定系数相近,都在0.30以上,而五种模型产生的相对均方跟误差分别为1.7241、1.9152、1.4308、1.9410和5.2731,与生长季相似,fao-radition模型模拟值偏大,不适合该地区非生长季草地蒸散发量的模拟,其他四个模型模拟值稍大于涡度计算值,作适当调整后可用于该研究区草地非生长季蒸散发量的模拟。调整后的penman、fao-penman、fao-penman-monteith和priestley-taylor四种模型模拟值与涡度计算值除去几个较大值外,较小值都比较集中的分布在直线y=x的附近,且调整后的四种种模型产生的相对均方跟误差都很小,分别为0.3624、0.3636、0.6820和0.3698,调整后四个模型和涡度数据可以作为研究区非生长季蒸散发量的观测,这对于研究区蒸散发的研究具有重要意义。4.FAO改进后的作物系数计算方法适合该区域草地作物系数的计算;FAO Penman-Monteith模型估算的日蒸散发量在0.50-7.26mm/d之间,生长季日均蒸散发量有年际变化,2011年>2014年>2012年>2013年,总体来看,土壤蒸发总量年际变化不大,影响蒸散发年际变化的主要部分是植被的蒸腾。5.通过对典型晴天下小时尺度实际蒸散发量与蒸发皿蒸散发量的关系的分析,将蒸发皿系数分为8:00-14:00和14:00-20:00两个时间段和8:00-20:00整个时间段来计算,8:00-14:00和14:00-20:00时间段的蒸发皿系数分别大于1和小于1,而8:00-20:00整个时间段蒸发皿系数小于1。更多还原

【Abstract】 Grassland accounts for about 25% of the world’s land area, is an important part of the ecosystem, and the ratio of grassland area of total area of Q ilian National Nature Reserve reached 50.54%, the subalpine meadow distributed widely and is the main part of the grassland ecosystem, which plays an important role in ecosystem service function in Q ilian Mountain area. The study area as water conservation forest has the most abundant vegetation and complete biological communities, which is located in the depths of the Q ilian Mountain. Subalpine meadow accounts for 7.45% of the total area of the study area, the ecological environment is very overgrazing, rodent activity is frequent, resulting fact the evapotranspiration of the grassland. Therefore, the study of subalpine meadow evapotranspiration on the local grazing has important guiding significance, also provide some reference for local authorities.The study using self- made lysimeter observe the evapotranspiration of subalpine meadow in the study area, ma inly includes: 1. observation of evapotranspiration at 8 am and 8 pm every day(including water evaporation evaporation); 2. observation of evapotranspiration on hour scale once a week(if weather conditions is bad, such as rainfall, push back a day, including water evaporation observation); 3. the relationship between evaporation and meteorological factor; 4. the characteristics of evapotranspiration obserced by lysimeter and calculated with vorticity data and its relationship with model simulation values; 5. with improved crop coefficient calculation method to calculate the coefficient of subalpine meadow, and tests this with the observed value. Draw the following conclusion:1. On the month scale, the order of the the average daily evapotranspiration is 2013>2011>2014>2012;In the year 2011 and 2012, the order of the the average daily evapotranspiratio of every month is June>August>July, and in the year 2013 and 2014, the order respectively are August >June>July and July>August>June. The average daily evapotranspiration has a certain relationship with the rainfall, the time of rain in one day is different, have different effects on evapotranspiration. The relationship between daily evapotranspiration and meteorological elements is wind speed>saturated vapor pressure>relative humidity>net radiation, the R2 is 0.4174, 0.3598, 0.3461 and 0.3322, all four meteorological elements have affect on daily evapotranspiration. The evapotranspiration of non- growing season(from the middle of October to mid-april the following year) calculated with vorticity data from September 16, 2013 to September 6, 2014, overall showed a trend of fluctuations, is small, basic below 1 mm/d, average daily evapotranspiration is 0.64 mm/d, the growing season(from the middle of October to mid-april) probably started in midapril, and the evapotranspiration began to increase, basically reaches the maximum on June, July and August, and then begin a stable state of fluctuation, the average daily evapotranspiration of growth season is 2.98 mm/day. The evapotranspiration and precipitation is very small on the non- growing season and is very large on growing season, so there may be a certain relationship between evapotranspiration and precipitation.2. On the evapotranspiration observed by the lys imeter, FAO-Penman-Monteith model is the best on day scale, and other three models(i.e., Penman, FAO-Penman, and Priestley-Taylor) are also good, FAO-Radiation model is bad; Priestley-Taylor model is the best one on hour scale, other four models are not s uitable for estimating hourly evapotranspiration in the grassland of the study area, the order of the input parameters significantly influencing evapotranspiration is radiation > air pressure > air humidity > air temperature>wind speed. When the radiation changes ±10%, the simulation values change about 20% and when other parameters change ±10%, the simulation values change less than 8%. Therefore, the radiation is the most sensitive factor.3. On the evapotranspiration calculated with vorticity data of growing season, the simulation value of FAO-Radiation model is bad and the simulation value other four models is light larger than the evapotranspiration calculated with vorticity data, so appropriate adjustments is made for the four models to simulate the evapotranspiration. And on the non- growing season, it is same to the growing season, the simulation value of FAO-Radiation model is bad and the simulation value other four models is light larger than the evapotranspiration calculated with vorticity data, so appropriate adjustments is also made for the four models to simulate the evapotranspiration. The simulation of evapotranspiration of non- growing season has a great significance for the study area.4. The improved crop coefficient formula recommended by FAO is suitable to be used in the study area, the daily ET0 estimated by,FAO Penman-Monteith model ranges from 0.50 mm/d to 7.26 mm/d, the ET0 has interannual difference and the sequence of ET0 in the study period is 2011 > 2014 > 2012 > 2013. From measurement data, we find soil evaporation having a little interannual change, that means the transpiration of grass is a key component to control interannual different of ET0. From the study, we can draw a conclusion that the actual evaportranspiration of grass in the subalpine meadow can be estimated by crop coefficients and reference evapotranspiration, both depending on weather data which are easily obtained by automatic weather station.5. Analyse the relationship between hourly actual evapotranspiration and pan evapotranspiration based on a typical sunny day, pan coefficient can be divided into 8:00-14:00 and 14:00-20:00 two periods and the whole time 8:00-20:00 to calculate, the pan coefficient of 8:00-14:00 and 14:00-20:00 is greater than 1 and less than 1, and the pan coefficient of 8:00-20:00 is less than 1.更多还原