Introduction

China has accounted for nearly one-fourth of the total rice-sown area and more than one-third of total rice production in the world since 1970 (State Environmental Protection Administration 2003). Rice is the staple food for more than 65 % of the Chinese population, accounting for 40 % of caloric intake in China (State Environmental Protection Administration 2003). However, rice crop suffers frequently from extreme weather events such as droughts, floods, cold and heat stress, which cause quite a large inter-annual variability in regional rice production (e.g., Xiang and Griffiths 1988; Jiao et al. 2008). Due to the increasing climate extreme events in recent decades under ongoing climate change (Zhai et al. 2005; IPCC 2007a), the risk of rice production to climate change and variability has increasingly been of concern (e.g., Tao et al. 2006, 2008a, b; Zhang et al. 2008; Li et al. 2009; Xiong et al. 2009; Tao and Zhang 2012). Nevertheless, the risk of climate variability or extreme weather events has often been addressed weakly, in comparison with those of mean climate change impacts. In fact, many of the deleterious impacts of global climate change could result from extreme and severe weather events such as extreme drought, heavy rainfall and extreme temperatures, rather than from changes in mean values of the atmospheric variables such as temperature (Salinger et al. 2000; IPCC 2007b). For an example, peaks of high temperature during the critical flowering period of a crop, even when occurring for just a few hours, could drastically reduce the production of important food crops, independently of any substantial changes in mean temperature (e.g., Matsui and Horie 1992; Wheeler et al. 2000).

Previous studies have indicated that the frequency and characteristics of weather and climate extremes are notably affected by ongoing climate change in recent decades (e.g., Moonen et al. 2002; Zhai and Pan 2003; Zhai et al. 2005; Alexander et al. 2006; IPCC 2007a; Piao et al. 2010; Vergni and Todisco 2010; Liu et al. 2011). These studies showed that since 1950 the number of heat stress had increased and widespread increases had occurred in the numbers of warm nights. The extent of regions affected by droughts has also increased as precipitation over land has marginally decreased while evaporation has increased due to warmer conditions (Tao et al. 2003). Generally, numbers of heavy daily precipitation events that lead to flooding have increased, but not everywhere (IPCC 2007a). In China, cold extremes have decreased while warm extremes have increased during the recent 50 years (Zhai and Pan 2003). The number of frost days in China has been reduced significantly. Although there is no significant trend in mean annual precipitation for China as whole, there exist distinct regional and seasonal differences. Significant increases in extreme precipitation have been found in western China, the mid-lower reaches of the Yangtze River, and parts of the southwest and south China coastal area (Zhai et al. 2005; Piao et al. 2010).

The changes in the frequency and characteristics of weather and climate extremes could have important implications for agricultural disasters and production (e.g., Tao et al. 2004; IPCC 2007b; Wang et al. 2008; Craufurd and Wheeler 2009; Li et al. 2009). However, so far, we have little knowledge on the characteristics of agricultural disasters changes, as well as the meteorological and agronomic causes, which are crucial for agriculture to cope with climate risk both at present and in future (Salinger et al. 2000; Stigter et al. 2000; Craufurd and Wheeler 2009; Tao and Zhang 2012). In the present study, we used both the observed records on rice disasters at agro-meteorological stations across China from 1991 to 2009 and the meteorological indexes to investigate the changes in rice disasters and climate extremes in recent decades. The objectives of this study were (1) to investigate the temporal and spatial changes in major rice disasters at station level across China in recent decades under ongoing climate change, (2) to understand the meteorological causes for rice disasters changes by investigating the changes in rice disasters-related climate extremes, and (3) to understand the characteristics of major rice disasters change in recent decades under the impacts of climatic and agronomic factors, and provide information to manage the disasters timely and scientifically.

Materials and methods

Observed datasets on weather and major rice disasters

Observed records on daily weather (including mean temperature, maximum and minimum temperature, and precipitation), rice phenology and disasters from 1991 to 2009 were obtained from national agro-meteorological stations of China, which were maintained by Chinese Meteorological Administration (CMA). Major crop disasters such as droughts, floods, cold and heat stress, and insects and diseases were discerned and recorded by local agronomists at the agro-meteorological stations of China. The records were gathered periodically, further compiled and shared by China meteorological data sharing service system of CMA (http://cdc.cma.gov.cn/). The records of the dataset included the station number, station name, province name, latitude, longitude, year, month, types of disaster, types of crop suffered from disaster, the month and date of disaster occurrence, degree of disaster (slight, moderate and severe), intensity of disaster (number of continuous days), area suffered from the disaster (ha), percentage of area suffered from the disaster (%). In the present study, the records from 1991 to 2009 at the agro-meteorological stations across the rice major cultivation regions in China were used in analyses. Single rice cropping system is dominant in northern China and northern part of the middle and lower reaches of the Yangtze River. Double rice cropping system, that is, rotation of early rice and later rice, is dominant in southern China.

Meteorological indexes for rice disasters

Besides the observed records on major rice disasters, we also used the meteorological indexes to investigate rice disasters-related climate extremes and their changes, and to understand to what extent the occurrence and change in the observed disasters could be ascribed to climate change and extremes.

Total precipitation and its standard deviation during rice growing period (from transplanting to maturity) were used as the meteorological indexes for rice droughts and floods disasters.

For heat stress, we focused on the middle and lower reaches of the Yangtze River, a major rice production region, where rice is vulnerable to heat stress. A meteorological index of high temperature was used on the basis of previous studies (e.g., Jin et al. 2009; Ren et al. 2010; Bao et al. 2012), which was documented to be fairly good in describing high-temperature stress on crop yield. The index counted the frequency of high-temperature events, which included the continuous high-temperature events [with daily mean temperature ≥30 °C for continuous 3 days or more in summer (June–July–August)] and the short episodes of high temperature (with daily maximum temperature ≥35 °C for continuous 3 days or more in summer).

Two types of injuries occurred due to low temperature, which were the delayed cool injury and sterile-type chilling injury (e.g., Sthapit 1992; Mavi and Tupper 2004). In the delayed cool injury, rice yields decreased due to insufficient grain maturation caused by low temperatures during the ripening period; or insufficient time was available for the grains to ripen fully before frost occurs in autumn, when flowering was delayed by low temperatures at a certain stage before heading. In the sterile-type chilling injury, rice yields decreased due to sterile spikelets caused by low temperatures at the booting or anthesis stage. The observed injury in developmental order was a stoppage of anther development, pollen unripeness, partial or no dehiscence, pollen grains remaining in anther loculi, little or no shedding of pollen grains on the stigma, and/or failure of germination on the stigma (Mavi and Tupper 2004). For the delayed cool injury, we focused on northeastern China, one of major rice production areas, where rice was vulnerable to delayed cool injury. We used the anomaly of temperature summation from May to September at a certain year (∆T), relative to its mean value during 1971–2000 (T), as the meteorological index (Table 1) (CMA 2009). As shown in Table 1, as an example, at a station where the mean temperature summation from May to September during 1971–2000 was between 80 and 85 °C, rice could suffer from common delayed cool injury if ∆T were between −1.3 and −1.1 °C at a certain year and could suffer from severe delayed cool injury if ∆T were between −2.6 and −2.2 °C at a certain year. For sterile-type chilling injury, the number of continuous days with daily minimum temperature or mean temperature less than a threshold value during a specific development stage was used as the meteorological index (Table 2) (CMA 2009). As shown in Table 2, as an example, in the middle and lower reaches of the Yangtze River and the south to them, rice could suffer from slight sterile-type chilling injury if the number of continuous days with daily minimum temperature ≤17 °C was 2 days at booting stage, and could suffer from severe sterile-type chilling injury if the number of continuous days with daily minimum temperature ≤17 °C was ≥5 days at booting stage.

Table 1 Meteorological index system for disaster grade of rice delayed cool injury in northeastern China (CMA 2009)
Full size table
Table 2 Meteorological index system for disaster grade of rice sterile-type chilling injury (CMA 2009)
Full size table

Statistics and analyses

By 2009, it was reported that 2000–2009 was the warmest decade on record since the beginning of modern instrumental measurements around 1850 (WMO 2010). Using both the observed records on major rice disasters and the corresponding meteorological indexes, we investigated the temporal and spatial changes in major rice disasters occurrence frequency during the warmer decade, in comparison with the period of 1991–2000. We further analyzed the relations between the changes in major rice disasters and ongoing climate change, as well as their meteorological and agronomic causes.

Results

Observed changes in major rice disasters

Droughts

During the period of 1991–2000, for early rice, droughts occurred mainly during transplanting and seedling stage in eastern China and before milky ripe stage in southern China (Fig. 1a; Table 3). During the period of 2000–2009, the stations suffered from droughts increased obviously in southeastern and southwestern China, and droughts occurred mostly during milky ripe and maturity stage (Fig. 1c; Table 3). Droughts frequency increased substantially in the major early rice cultivation area, although decreased in southern coast area (Fig. 1e).

Fig. 1
figure 1

Droughts occurrence frequency in rice growth stages during the period of 1991–2000 (a, b) and the period of 2000–2009 (c, d) based on the observed records at agro-meteorological stations, as well as the difference in total frequency over rice growing period between 2000–2009 and 1991–2000 (e, f), for early rice (a, c, e) and single or late rice (b, d, f), respectively. The size of symbol is positively proportional to the occurrence frequency

Full size image
Table 3 Major rice disasters occurrence frequency (times) in different rice growth stages during 1991–2000 and 2000–2009 based on the observed records at the rice agro-meteorological stations across China
Full size table

During the period of 1991–2000, single or late rice suffered from extensive droughts from tillering to maturity stage in southeastern China and from tillering to heading stage in northeastern China (Fig. 1b; Table 3). During the period of 2000–2009, the stations suffering from droughts increased in northeastern China, while decreased in southwestern China (Fig. 1d; Table 3). Droughts frequency increased substantially in southeastern and northeastern China, although decreased in portions of eastern and southwestern China (Fig. 1f).

Floods

During the period of 1991–2000, for early rice, floods occurred mainly from booting to maturity stage in eastern China, and from tillering to maturity stage in southern China (Fig. 2a; Table 3). In comparison with the period of 1991–2000, during the period of 2000–2009, stations suffering from floods increased in southern China, although decreased in eastern China; and floods occurred more frequently at earlier stages of rice development, for example, at the tillering and booting stage (Fig. 2c; Table 3). Floods frequency increased substantially in southern China, while decreased generally in portions of eastern China (Fig. 2e).

Fig. 2
figure 2

Floods occurrence frequency in rice growth stages during the period of 1991–2000 (a, b) and the period of 2000–2009 (c, d) based on the observed records at agro-meteorological stations, as well as the difference in total frequency over rice growing period between 2000–2009 and 1991–2000 (e, f), for early rice (a, c, e) and single or late rice (b, d, f), respectively. The size of symbol is positively proportional to the occurrence frequency

Full size image

During the period of 1991–2000, single or late rice suffered from floods generally before milky ripe stage (Fig. 2b; Table 3). During the period of 2000–2009, floods occurred more frequently during the tillering and booting stages, and the stations suffering from floods increased in northeastern and southwestern China (Fig. 2d; Table 3). Floods frequency increased substantially across their major cultivation areas, although reduced in portions of southern China (Fig. 2f).

Heat stress

During the period of 1991–2000, early rice suffered from heat stress in southeastern China mainly from booting to maturity stage (Fig. 3a; Table 3). During the period of 2000–2009, the stations suffered from heat stress increased in southwestern China (Fig. 3c; Table 3); moreover, heat stress occurrence frequency increased substantially in southeastern China (Fig. 3e).

Fig. 3
figure 3

Heat stress occurrence frequency in rice growth stages during the period of 1991–2000 (a, b) and the period of 2000–2009 (c, d) based on the observed records at agro-meteorological stations, as well as the difference in total frequency over rice growing period between 2000–2009 and 1991–2000 (e, f), for early rice (a, c, e) and single or late rice (b, d, f), respectively. The size of symbol is positively proportional to the occurrence frequency

Full size image

During the period of 1991–2000, single or late rice suffered heat stress in southern China generally before milky ripe stage (Fig. 3b; Table 3). During the period of 2000–2009, heat stress also occurred before milky ripe stage generally (Fig. 3d; Table 3); however, the frequency increased substantially in southeastern China and decreased in portions of eastern and southwestern China (Fig. 3f).

Chilling damage

During the period of 1991–2000, early rice suffered from chilling damage in southeastern China mainly from transplanting to booting stage (Fig. 4a; Table 3). During the period of 2000–2009, although also occurred from transplanting to booting stage (Fig. 4c; Table 3), chilling damage occurrence frequency increased substantially in southern and decreased in eastern China (Fig. 4e).

Fig. 4
figure 4

Chilling damage occurrence frequency in rice growth stages during the period of 1991–2000 (a, b) and the period of 2000–2009 (c, d) based on the observed records at agro-meteorological stations, as well as the difference in total frequency over rice growing period between 2000–2009 and 1991–2000 (e, f), for early rice (a, c, e) and single or late rice (b, d, f), respectively. The size of symbol is positively proportional to the occurrence frequency

Full size image

During the period of 1991–2000, single or late rice suffered from chilling damage disaster mainly from seedling to maturity stage in northeastern China, from heading to maturity stage in eastern China, and from transplanting to maturity stage in southwestern China (Fig. 4b; Table 3). During the period of 2000–2009, chilling damage occurred more often at later growth stage of rice in southwestern China (Fig. 4d; Table 3), in comparison with the period of 1991–2000. In addition, the frequency increased substantially in southwestern and northeastern China, although reduced in portions of southeastern and southwestern China (Fig. 4f).

Changes in climate extremes related to the major rice disasters

The disasters did not happen by accident but were the products of patterns of wind, temperature, and precipitation that produce climate extremes. Here, we used the meteorological indexes to investigate the related climate extremes and their changes to understand the meteorological causes for rice disasters changes.

Precipitation during rice growing period

During the period of 1991–2000, mean precipitation during early rice growing period (from transplanting to maturity) was generally above 300 mm in southern China (Fig. 5a). During the period of 2000–2009, mean precipitation during early rice growing period was also above 300 mm generally (Fig. 5c), however, the mean precipitation decreased by more than 100 mm at some stations and increased at others (Fig. 5e). The standard deviation of precipitation generally decreased at the stations with mean precipitation decreased, or vice versa (Fig. 5g).

Fig. 5
figure 5

Mean precipitation over rice growing period during the period of 1991–2000 (a, b) and the period of 2000–2009 (c, d), as well as the differences in mean precipitation (e, f) and standard deviation (STD) (g, h) between the two periods, for early rice (a, c, e, g) and single or late rice (b, d, f, h), respectively

Full size image

During the period of 1991–2000, mean precipitation during single or late rice growing period was generally above 300 mm in northeastern and southern China (Fig. 5b). However, during the period of 2000–2009, the precipitation became <300 mm at more stations in southeastern China (Fig. 5d). The mean precipitation decreased by about 100 mm at many stations in northeastern and southern China, although increased by around 100 mm at some stations in southwestern and eastern China (Fig. 5f). Similarly, the standard deviation of precipitation generally decreased at the stations with mean precipitation decreased, or vice versa (Fig. 5h).

The spatial patterns of precipitation change were well consistent with those of rice droughts and floods (Figs. 1, 2, 5). In comparison with the period of 1991–2000, for early rice, droughts increased (decreased) at 29 (10) stations during 2000–2009 (Fig. 1e), of which mean precipitation decreased (increased) at 27 (7) stations (Fig. 5e); floods increased (decreased) at 22 (25) stations during 2000–2009 (Fig. 2e), of which mean precipitation increased (decreased) at 8 (21) stations (Fig. 5e). For single or late rice, droughts increased (decreased) at 58 (67) stations during 2000–2009 (Fig. 1f), of which mean precipitation decreased (increased) at 47 (30) stations (Fig. 5f); floods increased (decreased) at 51 (27) stations (Fig. 2f), of which mean precipitation increased (decreased) at 27 (19) stations (Fig. 5f). The results suggested changes in rice droughts and floods should be largely attributed to changes in precipitation. In comparison with the period of 1991–2000, rice droughts frequency generally increased at the areas where mean precipitation decreased during the period of 2000–2009; and rice floods frequency generally increased at the areas where precipitation or/and its standard deviation increased during the period of 2000–2009.

Extreme high-temperature events

Based on the meteorological index, during the period of 1991–2000, high-temperature events occurred generally <20 times in northern part of the middle and lower reaches of the Yangtze River (Fig. 6a) and, however, generally more than 20 times in southern part. During the period of 2000–2009, high-temperature events occurred generally more than 20 times across all the 6 provinces in the middle and lower reaches of the Yangtze River (Fig. 6b). The frequency of high-temperature events increased substantially across the region (Fig. 6c). Daily mean temperature in summer increased by more than 0.5 °C between the two periods at many stations across the region (Fig. 6d). The standard deviation of daily mean temperature also increased except portions of eastern China (Fig. 6e), suggesting extreme high-temperature increase there.

Fig. 6
figure 6

High-temperature events occurrence frequency during the period of 1991–2000 (a) and the period of 2000–2009 (b), as well as the differences in occurrence frequency (c), summer mean temperature (d) and its standard deviation (STD) (e) between the two periods

Full size image

In the middle and lower reaches of the Yangtze River, according to the observed records, for early rice, heat stress frequency increased at 9 stations, decreased at 6 stations and remained unchanged during 2000–2009 (Fig. 3e), in comparison with the period of 1991–2000. For single and late rice, heat stress frequency increased at 5 stations, decreased at 7 stations and remained unchanged at one station (Fig. 3f). The changes in high-temperature events, as calculated by the meteorological index, represented well the changes of rice heat stress at all the stations with heat stress frequency increased, and at 2 stations with heat stress frequency decreased (Figs. 3e, f, 6).

Extreme low temperature events

Using the anomaly of temperature sum from May to September as a meteorological index for delayed cool injury (Table 1), we found the delayed cool injury occurred by a frequency of 10–30 % during the period of 1991–2000 in northeastern China (Fig. 7a). In contrast, during the period of 2000–2009, the occurrence frequency reduced substantially (Fig. 7b, c), because mean temperature from May to September increased (Fig. 7d) and its standard deviation decreased generally across northeastern China (Fig. 7e).

Fig. 7
figure 7

Rice delayed cool injury occurrence frequency in northeastern China based on meteorological index during the period of 1991–2000 (a) and the period of 2000–2009 (b), as well as the differences in occurrence frequency (c), mean temperature of May–September (d) and its standard deviation (STD) (e) between the two periods

Full size image

Using daily minimum temperature and the continuous days around rice booting, heading and flowering stage as a meteorological index for sterile-type chilling injury (Table 2), for early rice, the sterile-type chilling injury occurred generally by a frequency of 1–10 times during the period of 1991–2000 (Fig. 8a). During the period of 2000–2009, the sterile-type chilling injury occurred generally by a frequency of 1–3 times (Fig. 8c). Although generally decreased in southeastern China, the frequency increased in southwestern China (Fig. 8e). Daily mean temperature around rice booting, heading and flowering stage increased generally, although decreased at some stations in southern China (Fig. 8g), and the standard deviation generally increased in southeastern China and decreased in southwestern China (Fig. 8i). For single or late rice, the sterile-type chilling injury occurred by a frequency of 1–10 or more times during the period of 1991–2000 (Fig. 8b). The frequency of this type of injury decreased in southeastern China and, however, increased in northeastern and southwestern China, during the period of 2000–2009 (Fig. 8d, f). The daily mean temperature around rice booting, heading and flowering stage increased in southeastern and, however, decreased in northeastern and southwestern China (Fig. 8h), and the standard deviation changed reversely (Fig. 8j). The risk of chilling damage to rice increased in northeastern and southwestern China due to the increased extreme cold weather around rice heading dates during the warmer decade 2000–2009.

Fig. 8
figure 8figure 8

Rice sterile-type chilling injury occurrence frequency based on meteorological index during the period of 1991–2000 (a, b) and the period of 2000–2009 (c, d), as well as the differences in occurrence frequency (e, f), mean temperature from booting to flowering stage (g, h) and its standard deviation (STD) (i, j) between the two periods, for early rice (a, c, e, g, i) and single or late rice (b, d, f, h, j), respectively

Full size image

The resultant chilling damage changes based on the meteorological indexes were generally consistent with those based on the observed records at agro-meteorological stations (Figs. 4, 9). According to the observed records, for early rice, chilling damage frequency increased at 24 stations and decreased at 17 stations during 2000–2009 (Fig. 4e), in comparison with the period of 1991–2000. The meteorological index represented well the changes of chilling damage at 16 stations with chilling damage frequency decreased (Fig. 8e). For single and late rice, chilling damage frequency increased at 35 stations, decreased at 21 stations and remained unchanged at 3 stations (Fig. 4f). The meteorological index represented well the changes of chilling damage at 15 stations with chilling damage frequency increased and at 17 stations with chilling damage frequency decreased (Figs. 7c, 8f). The increases in rice chilling damage frequency at some stations in southern China were ascribed to decrease in mean temperature and increase in temperature variability (Fig. 8).

Fig. 9
figure 9

Insects and diseases occurrence frequency in rice growth stages during the period of 1991–2000 (a, b) and the period of 2000–2009 (c, d) based on the observed records at agro-meteorological stations, as well as the difference in total frequency over rice growing period between 2000–2009 and 1991–2000 (e, f), for early rice (a, c, e) and single or late rice (b, d, f), respectively. The size of symbol is positively proportional to the occurrence frequency

Full size image

Discussion

Climate change, climate extremes and disasters on rice production

The observed major meteorological disasters on rice production had obvious changes in recent decades. The spatial patterns of the disasters changes were generally consistent with those of corresponding meteorological indexes, suggesting changes in meteorological disasters in recent decades should be largely attributed to changes in climate extremes. For rice droughts and floods, both regions with increased and decreased disasters were identified, which were closely related to not only changes in precipitation but also changes in its standard deviation. The increased heat stress frequency at the investigated stations was ascribed to the increased extreme high-temperature events. For rice chilling damage, the meteorological index suggested the delayed cool injury should decrease substantially due to climate warming. Both observed records and the meteorological index indicated that the sterile-type chilling injury increased substantially at large areas due to increased climate extremes or/and decreased temperature, although decreased in some other stations. Therefore, the risk of chilling damage to rice production could increase at a large areas although climate becomes warming, since extreme weather could be more frequent due to the instability of the atmosphere (IPCC 2007a).

Besides the direct effects of climate change and climate extremes on agricultural meteorological disasters mentioned above, climate change could also contribute to changes in insects and diseases disasters of rice crop (Harvell et al. 2002). According to the observed records, during the period of 1991–2000, early rice suffered from insects and diseases in southeastern China mainly from booting to milky ripe stage (Fig. 9a). In contrast, during the period of 2000–2009, insects and diseases occurred more often at the earlier stage of rice development, for example, from seedling to booting stage (Fig. 9c). In addition, insects and diseases occurrence frequency reduced in many stations in southeastern China, although increased in some other stations (Fig. 9e). During the period of 1991–2000, single or late rice suffered from insects and diseases from seedling to maturity stage, particularly from booting to heading stage (Fig. 9b). In contrast, during the period of 2000–2009, insects and diseases occurred more often at the earlier stage of rice development (Fig. 9d). In addition, the occurrence frequency decreased in the central part of southern China, although increased in portions of eastern China and southwestern China (Fig. 9f).

Although rice production was documented to be benefited from long-term mean climate trends in the past few decades in large areas of China (Tao et al. 2008a), the risk to climate extremes of rice production was increasing with ongoing climate change (Tao and Zhang 2012). Climate extremes and the related disasters in the past few decades can cause quite a large loss on rice production (e.g., Simelton 2011).

Phenology and cultivars change, agronomic practices and the implications for disaster reductions

The meteorological indexes used in the study, which were based on many previous studies, represented explicitly climate extremes events. However, the agronomic management practices and the sensitivity of varieties were not taken into account by the meteorological indexes. The inconsistence between the changes in observed disasters records and the changes in the meteorological indexes at some stations could be largely due to the complexity of disasters occurrence, the roles of agronomic management practices and varieties characteristics. For example, previous studies showed that temperatures >35 °C for more than 1 h could lead to pollen sterility in rice (Viraktamath and Virmani 2001). Some studies showed that air temperature per se was not sufficient to predict the occurrence of heat-induced sterility, but factors influencing the heat budget of the panicles were needed to account for the crop damages under open field conditions, which could lead to higher or lower panicle temperature than air temperature (Hasegawa et al. 2009; Yoshimoto et al. 2009). Yoshimoto et al. (2009) suggested that the panicle temperature instead of air temperature should be used as a measure variable for the climate change impact study, especially on rice sterility. Furthermore, heat stress damage was also subjected to the variety sensitivity and management practices of the crop (Hasegawa et al. 2009). The decrease in heat stress at some stations could be ascribed to changes of varieties and agronomic management practices.

Likewise, the occurrence frequency of rice insects and disease increased at some stations and decreased at others, although some previous studies showed that climate warming would affect host–pathogen interactions by increasing pathogen development rates, transmission and the number of generations per year, relaxing overwintering restrictions on pathogen life cycles and modifying host susceptibility to infection (Harvell et al. 2002). Rice insects and diseases were subjected to type of insects, host plants and management practices, besides climate variables such as temperature, precipitation and wind. These results suggested although rice disasters in recent decades could be largely attributed to changes in climate change and climate extremes, sometimes the links between these disasters with climate changes and climate extremes were complex due to the multivariate nature of climate change, the nonlinear thresholds in the disasters and climate processes, as well as the roles of varieties changes and agronomic management practices.

The changes in agricultural disasters could be reduced by agricultural management practices, such as the adoption of heat-tolerant cultivars to prevent from heat stress and the use of pesticides to control insects and diseases. In the study, we found floods, insects and diseases occurred more frequently at earlier stage of rice development during the period of 2000–2009, in comparison with the period of 1991–2000. These changes could be ascribed to climate change and consequently the advance of rice and insects phenology (Tao et al. 2006; Zhang et al. submitted). In contrast, chilling damage occurred more frequently at later growth stages of rice in southwestern China. Besides the changes in climate extremes, these changes could be ascribed to the adoption of late mature cultivar by farmers to take advantage of increased heat resources from climate change, which was, however, subjected to higher climate variability (Zhang et al. submitted). These findings suggested that options should be taken timely and scientifically to reduce the disasters and to cope with ongoing climate change.

Conclusion

In the study, we used both the observed records and the meteorological indexes to investigate the major rice disasters and their changes in recent decades. We analyzed the temporal and spatial changes in the major rice disasters occurrence frequency and their relationships to climate change, climate extremes and crop phenology. In comparison with the period of 1990–2000, during the warmer decade 2000–2009, rice droughts frequency increased substantially in southeastern and northeastern China, and decreased in eastern and southwestern China. Floods frequency increased substantially across the cultivation area and reduced in portions of southeastern China. The heat stress occurrence frequency increased substantially in southeastern China and decreased in some locations in eastern and southwestern China. The chilling damage occurrence frequency increased substantially in southwestern and northeastern China, and decreased in portions of southeastern and southwestern China.

The consistency between the patterns of observed disasters and those derived from meteorological indexes suggested changes in meteorological disasters in recent decades could be largely attributed to changes in climate extremes and closely related to ongoing climate change. Nevertheless, we found sometimes the links between these disasters with climate changes and climate extremes were complex due to the multivariate nature of climate change, and nonlinear thresholds in both the disasters and climate processes. Agricultural management practices could induce or reduce agricultural disasters.

We also found that floods, insects and diseases occurred more frequently at earlier growth stages during the period of 2000–2009, in comparison with the period of 1991–2000. In contrast, chilling damage occurred more frequently at later growth stages in southwestern China. Our findings suggested that options should be taken timely and scientifically to reduce the disasters and to cope with ongoing climate change, based on the new characteristics of agricultural disasters changes in recent decades.