Search

Rainfall, temperature, and solar radiation are important climate factors, which determine crop growth, development and yield from instantaneous to decadal scales. We propose to identify year patterns of climate impact on yield on the basis of rain and non-rain weather. There are inter-related impacts of climatic factors on crop production within a specific pattern. Historical wheat yield data in Queensland during 1889-2004 were used. The influence of meteorological conditions on wheat yields was derived from statistical yield data which were detrended by 9-year-smoothing averages to remove the effects of technological improvements on wheat yields over time. Climate affects crop growth and development differently over different growth stages. Therefore, we considered the climate effects at both vegetative and reproductive stages (before and after flowering date, respectively) on yield. Cluster analysis was employed to identify the year patterns of climate impact. Five patterns were significantly classified. Precipitation during the vegetative stage was the dominant and beneficial factor for wheat yields while increasing maximum temperature had a negative influence. Crop yields were strongly dependent on solar radiation under normal rainfall conditions. As the effect of rainfall on soil water is relatively long-lasting, its beneficial effect in vegetative stage was higher than its effect during the reproductive stage. The Agricultural Production Systems sIMulator ( APSIM) was evaluated using long-term historical data to determine whether the model could reasonably simulate effects of climate factors for each year pattern. The model provided good estimates of wheat yield when conditions resulted in medium yield levels, however, in extremely low or high yield years, corresponding to extremely low or high precipitation in the vegetative stage, the model tended to underestimate or overestimate. Under high growing season precipitation, simulations responded more favourably to reproductive stage rainfall than measured yields. [ABSTRACT FROM AUTHOR]

The Australian Alpine region is highly vulnerable to extreme climate events such as heavy rainfall and snow falls, these events subsequently impact rainfall erosivity and hillslope erosion in the region. In this study, the relationship between extreme rainfall indices (ERIs) and rainfall erosivity was examined across the Alpine region in New South Wales (NSW) and Australian Capital Territory (ACT) and the surrounding areas including Murray and Murrumbidgee and South East and Tablelands (SET). Rainfall erosivity, hillslope erosion, and their changes were estimated in the future periods using the revised universal soil loss equation and the NSW/ACT Regional Climate Modeling (NARCliM) projections. Results from the study demonstrate a good relationship between ERIs (especially Rx5Day) and rainfall erosivity. The rainfall erosivity and hillslope erosion are projected to increase about 2 and 8% for the near future (2020–2039), further increase to 8 and 18% for the far future (2060–2079) in the Alpine region assuming the groundcover is maintained at the current condition. The change in rainfall erosivity and erosion risk is highly uneven in space and in season with the highest erosion risk in summer with an increase about 33% in the next 50 years. The highest erosion risk area is predicted within SET (maximum rate 19.95 Mg ha−1 year−1), but on average, the ACT has the highest erosion rate, which is above 1.36 Mg ha−1 year−1 in all periods. The snowmelt in spring in the Alpine region is estimated to increase the rainfall erosivity by 13% in the baseline period, up to 24% in the near future, but far less (about 1%) in the far future due to predicted temperature rise and less snow available in the Alpine region in the next 50 years. [ABSTRACT FROM AUTHOR]

Drought is viewed as a naturally recurring phenomenon in many Australian agricultural systems. Identifying regional changes in frequency and severity of drought induced by climate change is required to develop regionally specific adaptation strategies. In this study, we provided a first look at the impacts of climate change on 21st century drought characteristics over the New South Wales wheat belt of southeastern Australia. These impacts were assessed from an ensemble of 28 statistical downscaled global climate models under representative concentration pathway (RCP8.5). A modified relative standardized precipitation and evapotranspiration index (rSPEI) at the seasonal scale (3 months) was used to analyse temporal and spatial changes in drought. Results indicated that there was a tendency towards more frequent and severe winter–spring droughts over the study area. Moreover, winter–spring drought prone areas were expected to expand from west to east. Until the end of the 21st century, more than half the wheat belt would be vulnerable to winter–spring drought. The combined effects of reduced precipitation and increased temperature during future winter and spring seasons were the main reasons causing these changes of drought. In addition, summer and autumn droughts would have both slight temporal and spatial changes across the study region. This study also revealed that traditionally dry areas would likely experience an increased frequency of drought compared to wetter areas when subjected to a same increase in temperature or decrease in precipitation. Furthermore, the western part of the wheat belt might be unsuitable for winter crops in the future, or at least exposed to an increased risk of variable yield and would require a gradual transformation which might include summer crops or pastures. Investments in cropping land should be focused on the east part of the wheat belt to achieve more consistent financial returns. Drought prone areas were identified in the study region using the modified drought index, rSPEI. Spring and winter drought prone areas were likely to expand remarkably. By the end of the 21st century, over half of the study area was at a high risk of suffering spring and winter drought. Summer and autumn drought prone areas were primarily in the southwest and northwest and might only experience slight changes. [ABSTRACT FROM AUTHOR]

Global climate models (GCMs) are useful tools for assessing climate change impacts on temperature and rainfall. Although climate data from various GCMs have been increasingly used in climate change impact studies, GCMs configurations and module characteristics vary from one to another. Therefore, it is crucial to assess different GCMs to confirm the extent to which they can reproduce the observed temperature and rainfall. Rather than assessing the interdependence of each GCM, the purpose of this study is to compare the capacity of four different multi‐model ensemble (MME) methods (random forest [RF], support vector machine [SVM], Bayesian model averaging [BMA] and the arithmetic ensemble mean [EM]) in reproducing observed monthly rainfall and temperature. Of these four methods, the RF and SVM demonstrated a significant improvement over EM and BMA in terms of performance criteria. The relative importance of each GCM based on the RF ensemble in reproducing rainfall and temperature could also be ranked. We compared the GCMs importance and Taylor skill score and found that their correlation was 0.95 for temperature and 0.54 for rainfall. Our results also demonstrated that the number of GCMs ensemble simulations could be reduced from 33 to 25 in RF model while maintaining predictive error less than 2%. Having such a representative subset of simulations could reduce computational costs for climate impact modelling and maintain the quality of ensemble at the same time. We conclude that machine learning MME could be efficient and useful with improved accuracy in reproducing historical climate variables. The relative change of RMSE for different size of the CMIP5 subset compared with full number of GCMs using the random forest (RF) ensemble simulations. Our results demonstrated that the number of GCMs ensemble simulations could be reduced from 33 to 25 in RF model while maintaining predictive error less than 2%. Having such a representative subset of simulations could reduce computational costs for climate impact modelling and maintain the quality of ensemble at the same time. [ABSTRACT FROM AUTHOR]