摘要
玉米是世界上种植最为广泛的谷类作物,我国绝大部分地区的气候和土壤条件都适合玉米种植,是世界上第二大玉米生产国家,拥有纵穿全国的7大玉米带。其中北方春播玉米区在我国玉米生产中占有重要地位,对解决我国粮食问题起到举足轻重的作用。我国气候复杂,时空变化幅度较大,气象灾害种类多、发生频率高。其中,干旱是造成玉米减产的主要气象灾害。辽宁省属于北方春玉米区,是我国玉米主产区之一。本文以辽宁省为研究区域,利用全省23个站点的气象、玉米产量和播种面积资料分析玉米干旱风险。主要工作为,运用概率统计和聚类分析方法,以水分亏缺率和减产率为衡量指标,分析辽宁省玉米全生育期水分亏缺和减产的发生强度、发生概率及其时空分布,并依据减产率进行减产区域划分;运用信息膨化的方法处理玉米全生育期内的气象要素值,对各气象因子与产量进行通径分析;以全生育期和水分关键期的水分亏缺率为因变量,构建玉米灾损回归模型;最后以干旱灾害发生的时间频率、干旱灾害强度、玉米暴露水平和玉米生产水平来构建玉米干旱风险分析指数,计算1961-2006年各站点的风险指数,据此进行玉米干旱风险区划及风险分析。主要结论如下:
1.通过分析1961-2006年气象产量的波动情况发现,减产年、平年和增产年交替出现,呈明显的周期性波动,平均每2-3年发生一次减产,其中正值年份占52%,负值年份占48%。根据各站点减产率的发生概率将20个站点分为A、B两个减产区和极端、严重、重度、中度、轻度五个减产年型。辽宁省西部地区即A区,玉米减产情况较重,严重以上减产年型发生概率较高,在45%以上;中东部和南部地区即B区,玉米减产情况较轻,主要是轻度和中度减产年,发生总概率达70%。从减产年型发生概率的年代变化来看,A、B两个区域均是在80和90年代减产最为严重,尤其是A区,90年代重度以上减产年型发生概率达70%;2000年后减产情况有所好转。A区内严重和极端减产年型在2000年后的发生概率均低于80和90年代,B区内除轻度减产年型外,其它减产年型发生概率均是在2000年后最低。
2.根据各站点全生育期水分亏缺率的分布概率和数值范围,将23个站点分为A、B、C、D四种概率分布类型和极端、重度、中度和轻度四个水分亏缺年型。A区极端水分亏缺年型出现概率最高,且在年代间呈起伏式发展;B区有极端水分亏缺年但发生概率较低;C区和D区没有出现极端水分亏缺年。水分亏缺在90年代最为严重。
3.对各站点气象产量与膨化处理的气象数据进行相关分析,结果发现,影响产量的主要气象因子为降水和最低气温。降水的影响主要发生在夏秋季节;最低温度的影响时段,1981年前在8月末到9月末,为负相关,而1981年之后,低温的影响几乎都发生在八月底之前即夏季,为正相关。通过对气象因子及作物潜在蒸散量与气象产量的通径分析可知,降水量和作物潜在蒸散量在不同年代段和区域内均为影响产量的主要因子。所以,以降水量和作物潜在蒸散量构建水分亏缺率干旱指标,根据减产率分析中所得到的A、B两个减产区,以全生育期和水分关键期的水分亏缺率为因变量与减产率构建灾损模型,A区域线性模拟效果较好,且模拟值与真实减产率差异性检验为不显著。B区域线性模拟效果不理想,模拟值与真实减产率差异性检验为显著,即模拟值与真实值差异很大,不能用来模拟真实减产率。但是,B区域中水分亏缺百分率与相对气象产量的相关系数均很高。
4.据标准化后玉米干旱风险指数的数值范围,将辽宁省划分为高等、中等、低风险三个风险区域。高风险区风险指数≥0.3,此区主要分布在辽宁西部干旱地区和南部的大连地区,玉米全生育水分亏缺年份发生概率接近60%,灌浆期水分亏缺发生概率大于55%,抽雄吐丝期最大水分亏缺率大于40%。加上玉米种植面积较大,综合干旱风险最大;中等风险区风险指数在0.1～0.3之间,主要分布在辽西走廊半干旱区和辽河下游平原地区,全生育期平均降水量基本能够满足玉米生长需求,水分亏缺年份发生概率仅13%,灌浆期水分亏缺发生概率很高,达到44%,玉米暴露风险和农业生产水平均较高,综合干旱风险为中等;低风险区风险指数<0.1,主要分布在辽东山区,全生育期、抽雄开花期和灌浆期降水量均很充足,能够满足玉米生长需求,全生育期水分亏缺发生概率小于10%,灌浆期和抽雄开花期水分亏缺发生概率为37%和16%,玉米种植面积小,生产水平较高,综合干旱风险最小。
5.为了验证玉米干旱风险指数,以高风险区为例,分析玉米干旱风险指数与实际减产率的相关性和回归性。结果发现,二者存在很强的负相关性,在大部分年份点高的风险指数对应高的减产率,相关系数为-0.39,线性回归分析伴随概率为0.01,达到了极显著水平(SPSS)。将风险指数和减产率序列中典型的低温冷害年份去掉后,再对二者做相关分析,相关系数达到了-0.7,玉米干旱风险指数与相对气象产量直线回归相关系数>0.5,风险指数较为准确。
Maize is the world's most widely planted cereal crops. The climate and soil conditions are suitable for planting maize in most areas of China. China is the world’s second-largest maize-produting country, and has 7 maize-produting regions. Northern Spring Maize Region is one important maize-produting region, and plays a decisive role in solving national food issue. The climate is very complexity with significant different between years in China. Meteorological disasters always affect maize production. And drought is one of main meteorological disasters. Liaoning Province is included in Northern Spring Maize Region, and is one of main producing areas in China. In this paper, take Liaoning Province as study area to analyze drought risk of maize, based on the data of meteorological and maize yield and planting area of more 23 sites throughout the province. Take yield reduction rate of maize and water deficit as indexs to analyze the intensity, occurrence probability, temporal and spatial variation of water deficit rate and yield reduction rate with the method of probability statistics and cluster analysis. And regionalize Liaonig Peovince into two regions based on yield reduction rate.Using the information expanding method to treat meteorological data. And then analyze the impact of meteorological factors on yield with path analysis method. Construct regression model of maize drought yield reduction rate and water deficit rate of the whole and key periods. Construct drought risk comprehensive index with indexs of occurrence frequency of drought, intensity of drought, exposure and production level of maize as evaluating. At last, analyze maize drought risk and regionalization based on the drought risk comprehensive index of every site from 1961 to 2006.The main conclusions are as follows:
1. The fluctuations of meteorological yield from 1961 to 2006 showed interactively fluctuation of negative and positive year with a cyclical occurred, and the yield reduction occured every 2-3 years. Positive year occupied 52% and negative year occupied 48%. According to the yield reduction rate of 20 sites to regionalize Liaoning Province into A and B tow ragions and extreme, serious, severe, moderate, and mild, five types of yield reduction year. Ragion A is in the western of Liaoning where the yield reduction was seriuous. And the occurance probability of yield reduction above serious intensity was more than 45%. Region B is in the eastern and southern parts of the province where yield reduction intensity was lighter. There mainly happened mild and moderate yield reduction, and the total occurrence probability was nearly 70%. View of the yield reduction changes by ages, the most severe yield reduction ages were 80’s and 90’s in two regions. Particularly in region A, the accurance probability of yield reduction above serious intensity was nearly 70%. The yield reduction became lighter after 2000. The accurance probability of serious and extreme yield reduction was lower after 2000 than that of 80’s and 90’s. The accurance probability of each yeild reduction year type was lower after 2000 than other ages except the light yield reduction year type.
2. According to the water deficits accurance probability and value range of each site diveided 23 sites into A, B, C, D four kinds of probability distribution type regions and extreme, severe, moderate and light four water deficit year stypes. Ragion A had the highest occurrence probability of extreme water deficit, and showed a choppy development between yeazrs. In Ragion B, there had extremely water deficit year with low occurrence probability. There was no extreme water deficit occured in Ragion C and D. And the occurrence probability of water deficit was highest in 90’s.
3. Through the correlation analysis of meteorological yield and expanding treated meteorological data of sites, we can see that the main meteorological factors effecting maize yield were precipitation and temperature. The main effecting periods of precipitation were in summer and autumn. The main effect period of minimum temperature was from late August to late September, before 1981, and the correlation was negative. After 1981, the effecting period of minimum almostly occurred in summer, before the end of August, and the correlation was positive. The result of path analysis about meteorological factors and potential evapotranspiration with meteorological yield indicated that the precipitation and potential evapotranspiration were main factors affecting yield whether in different perionds or different regions. So conduct water deficit rate with precipitation and potential evapotranspiration as drought index. Take water deficit rate of the whole growth period and key periods as dependent variables to construct regression models with yield reduction rate in two yield reduction regions. The result of simulating in Region A was better, and the diversity test of actual valve and similation value was not significant. The simulation result was not good in Region B, the two independent samples test for the difference was significant, that means the model can not be used to simulate the actual yelid reduction rate. But the correlation coefficents of yield reduction rate and water deficit rate were all very haigh.
4. According to the value range of standared maize drought risk index, devided Liaoning Province into high, moderate, low-risk, three risk regions. RegionⅠwas high-risk zone, where the risk index was more than 0.3. This region is mainly in the western arid area and Dalian. The occurance probability of water deficit in the whole growth period was close to 60%. The occurance probability of water deficit in filling stage was more than 55%. The highest value of water deficit rate in silking stage was more than 40%. Plus, the highest exposure risk of maize, the integrated drought risk was largest in there. RegionⅡwas moderate risk zone, where the risk index was between 0.1 to 0.3, mainly in the West Liaoning Corridor semi-arid area and downstream Plain of Liao River of the province. The mean precipitation of the whole growth period can basicly meet tthe needs, and the occurance probability of water deficit was only 13%. The highest occurance probility was in filling stage, which was close 44%. The product level and the exposure risk index were at a higher level, there. Totally, the integrated drought risk was moderate. RegionⅢwas low-risk zone, where the risk index was less than 0.1. This area mainly distributed in eastern mountainous areas of the province. There had abundant precipitation to meet the growth demand of mazie in the whole growth periord, silking stage and filling stage. The occurance probability of water deficit in the whole growth period was less than 10%. And the occurance probability of water deficit in silking and filling stage were 37% and 16%. In addition, the mazie planting area was small, and production level was high. So the integrated drought risk was smallest there.
5. Take high drought risk region as example to check the maize drought risk index, through analyze the correlation of actual yield reduction rate and drought risk index value. There was negative correlation between drought risk index and yield reduction rate at a certain degree. The high risk index point corresponds to high reduction rate in most years, and the correlation coefficient was -0.39. The result of linear regression analysis was significant with a probability of 0.01, which achieved a very significant level (SPSS). Reanalyze the correlation after removing the chilling damage years from the data series. The correlation was -0.7. And the linear regression correlation of drought risk index and yield reduction rate was more than 0.5. The drought risk index was validated to be reasonzble.
引文
1.陈锡康.全国粮食产量预测研究.中国科学院院刊, 1992(4):330～332.
2.程伟,刘国璧.灰色系统理论在粮食产量预测中的应用.湖南工程学院学报,2008, 18(3):64～67.
3.戴阁丈文,旭洲.辽宁农业气象干旱分析.辽宁气象,1996(1):29～30.
4.邸伟.关于辽宁玉米发展问题的探讨.农业经济,2008(6):16～17.
5.段一贯.辽宁省干旱分析.东北水利水电,1991(5):29～35.
6.郭庆法,王庆成,汪黎明.中国玉米栽培学.上海科技出版社,2004,9:32～33.
7.贺振,贺俊平.基于灰色理论的河南省粮食生产预测.中国农学通报,2008, 24(6):475～478.
8.黄崇福,刘新立,周国贤等.以历史灾情资料为依据的农业自然灾害风险评估方法.自然灾害学报,1998,5,7(2):1～8.
9.黄崇福.自然灾害风险评价理论与实践.北京:科学出版社,2005,5～15.
10.霍治国,李世奎,王素艳,刘锦銮,薛昌颖.主要农业气象灾害风险评估技术及其应用研究.自然资源学报,2003,11,18(6):692～701.
11.纪瑞鹏,班显秀,张淑杰.辽宁地区玉米作物系数的确定.农业环境科学,2004,20(3):246～248.
12.金菊良,魏一鸣,付强等.改进的层次分析法及其在自然灾害风险识别中的应用.自然灾害学报,2002,5,11(2):20～24.
13.李晶,张运福,班显秀.辽宁玉米作物系数研究.辽宁气象,2000(1):18～19.
14.梁泽学,江静.中国北方地区1961～2000年干旱半干旱化趋势.气象科学,2005,2,25(1):9～15.
15.刘布春,王石立,冯玉平.国外作物模型区域应用研究进展.气象科技,2002,30(4): 194～203.
16.刘布春,王石立,庄立伟等.基于东北玉米区域动力模型的低温冷害预报应用研究.应用气象学报,2003,14(5):616～624.
17.刘树泽,张宏铭,蓝鸿等.作物产量预报方法.气象出版社, 2007,1:67～179.
18.刘芝绅,朱海,王小博等.辽宁省玉米生产区域化布局研究.农业经济,2007 (9):58～59.
19.马树庆,王琪,罗新兰.基于分期播种的气候变化对东北地区玉米(Zeamays)生长发育和产量的影响.生态学报, 2008, 28(5): 2131～2139.
20.马树庆,袭祝香,王琪.中国东北地区玉米低温冷害风险评估研究.自然灾害学报,2003,12(3):138～140.
21.门福禄.关于灾害学和灾害研究方法若干问题的浅见.自然灾害学报, 2002,11(4):149～152.
22.商彦蕊.干旱、农业旱灾与农户旱灾脆弱性分析—以邢台县典型农户为例.自然灾害学报,2000,5,9(2):55～61.
23.尚红敏.几种冬小麦产量预报方法介绍.河南气象, 1998(1):36～37.
24.邵晓梅,许月卿,严昌荣.黄河流域气候水分盈亏时空格局分析.气候与环境研究,2007,12(1):75～78.
25.谈克俭,张旭.辽宁省玉米发展的远景分析.农业经济,2005(8):50.
26.谭凯琰,李郁竹.遥感植被指数与冬小麦叶面积系数和麦土比的关系.遥感信息,1989,1:63～67.
27.田庆九.植被指数研究进展.地球科学进展,1998,(4):56～61.
28.宛公展.华北地区冬小麦产量预报模式的初步研究.气象学报,1984,42(3):357～363.
29.王春乙,娄秀荣,王建林.中国农业气象灾害对作物产量的影响.自然灾害学报, 2007,16(5):37～38.
30.王道龙,钟秀丽,李茂松,杨修.20世纪90年代以来主要气象灾害对我国粮食生产的影响与减灾对策.灾害学,2006,21(1):18～19.
31.王石立,娄秀荣.华北地区冬小麦干旱风险评估的初步研究.自然灾害学报,1997,8,6(3):63～68.
32.温克刚等。中国气象灾害大典(辽宁卷).北京:气象出版社,2005.
33.谢安,孙永罡,白人海.中国东北近50年干旱发展及全球气候变暖的响应.地理学报,2003,58:75～81.
34.姚玉璧,张存杰,邓振镛,等.气象、农业干旱指标综述.干旱地区农业研究,2007,25(1):185～189.
35.尹东.农业产量预报技术的研究和应用.干旱气象,2007,6,25(2):12～16.
36.于武盛,王守杰,吕锦有等.辽宁省地表水资源分布及成因分析.农业科技与装备,2008(2):25～26.
37.袁文平,周广胜.干旱指标的理论分析与研究展望.地球科学进展, 2004, 19(6):982～988.
38.张继权,李宁.主要气象灾害评价与管理的数量化方法及其应用.北京:北京师范大学出版社,2007,20～35.
39.张雪芬,毛留喜,陈怀亮.河南省夏玉米农学模式产量预报[J].河南气象, 1996(2): 22～23.
40.张玉革,姜勇,王力.气候条件对玉米生产的影响.辽宁农业科学,1998(5):17～19.
41.赵春雨,曲晓波,王颖等.辽宁气候变化及若干气象灾害的事实分析.灾害学,2007,22(4):77～80.
42.赵聚宝,李克煌.干旱与农业干旱.中国农业出版社,1995,8:192～209.
43.赵先丽,李丽光,贾庆宇等.1988—2007年辽宁主要农业气象灾害分析.气象与环境学报,2009,25(2):33～37.
44.钟以章,蒋秀琴,陈爱萍.辽宁省自然灾害及基本特点.灾害学,1993,8(1):51～55.
45.朱海,刘芝绅,王小博.辽宁省玉米产业发展现状及对策.中国农业资源与区划,2006,27(5):28～31.
46.朱琳,叶殿秀,陈建文,郭兆夏,李美荣.陕西省冬小麦干旱风险分析及区划.应用气象学报,2002,4,13(2):201～206.
47. Bonaccorso B., Bordi I., Cancielliere A. et al. Spatial variability of drought: An analysis of the SPI in Sicily. Water Resources Management, 2003(17):273～296.
48. Bouman B AM, van Keulen, van LaarHH, et al. The School of de Wit crop growth simulation models: a pedigree and historical overview[J]. Agricultural Systems, 1996, 52(2～3): 171～198.
49. C. Emdad Haque. Perspectives of Natural Disasters in East and South Asia, and the Pacific Island States: Socio-economic Correlates and Needs Assessment. Natural Hazards 29: 465～483, 2003.
50. Denmead O. T., Shaw R. H. The effects of soil moisture stress at different stages of growth on the development and yield of corn, Agronomy Journal, 1960, 52:272～274.
51. Domonkos, P. Recent precipitation trends in Hungary in the context of larger scale climatic Changes.Natural Hazards, 2003(29):255～271.
52. Hayes M., Wilhite D. A., Svoboda M. et al. Monitoring the 1996 drought using the standardizedprecipitation index. Bulletin of the American Meteorological Society, 1999(80):429～438.
53. Hong Wu, Donald A., Wilhite. An Operational Agricultural Drought Risk Assessment Model for Nebraska, USA. Natural Hazards , 2004,33: 1～21.
54. Hoogenboom G, Wilkens P W ,Tsuji G Y (eds). University of Hawaii. Honolulu, Hawaii, 1999. 3(4): 28.
55. Kropff M J, van Laar H H and Matthews R. ORAZA1, An Ecophysiological Model for Irrigation Rice Production[J]. SARP Research Proceedings. DLO Research Institute for Agrobiology and Soil Fertility Wageningen, Netherlands, 1994: 110.
56. Lawson P M. Nebraska droughts: a study of their last chronological ang spatial extent with implications for the future.Dept .Geogr. university of Nebraska , 1971,1:82～87.
57. Lloyd, Hughes B., Saunders M. A. A drought climatology for Europe. International Journal of Climatology, 2002b(22):1571～1592.
58. McKee T. B., Doesken N. J., Kleist, J. The relationship of drought frequency and duration to time steps, Preprints, 8th Conference on Applied Climatology. Anaheim, California, 1993(17):179～184.
59. Meinke H, Hammer G L and Selvaraju R. Using seasonal climate forecasts in agriculture———the Australian experience[A]. In: Sivakumar M V K, eds. Climate Prediction and Agriculture. Proceedings of the START/WMO International Workshop held in Geneva, Switzerland, 27～29 September 1999. Washington, DC, USA, International START Secretariat, 2000,195～214.
60. Min S. K., Kwon W. T., Park E. H. et al. Spatial and temporal comparisons of droughts over Korea with East Asia. International Journal of Climatology, 2003(23):223～233.
61. Palmer, W. C. Department of Commerce Weather Bureau Research. Meteorological Droughts, U.S. ,1965, 45～58.
62. Penning de Vries F W T, Jansen D M, ten Berge H F M, et al. Simulation of ecophysiological processes of growth of several annual crops. Simulation Monographs. Centre for Agricultural Publishing and Documentation (Pudoc). Wageningen, Netherlands, 1989.
63. Rouault, M., Richard, Y. Intensity and spatial extension of droughts in South Africa at different time scales. Water SA ,2003(29):489～500.
64. Sergio M. Differences in Spatial Patterns of Drought on Different Time Scales: An Analysis of the Iberian Peninsula .Water Resources Management , 2006(20):37～60.
65. Starr T. B., Kostrow P. I. The response of spring wheat yield to anomalous climate sequences in the United States, Journal of Applied Meteorology. 1978, 17(8):1101～1115.
66. Ts. E. Mirtskhulava. Assessment and Management of Land Flooding Risk. Eurasian Soil Science, 2007, 40(11):1233～1241.
67. Tsakiris G., Vangelis H. Towards a drought watch system based on spatial SPI, Water Resources Management, 2004(18):1～12.