Session 1.1 Seasonal Forecasts and Climate Extremes
For a complete list of all of the workshop abstracts click here (PDF).
If you are presenting a poster, please view the Poster Specifications here
Session 1.1: Oral Presentation
Title: Past and future weather-induced risk in crop production
Author: Joshua Elliott
University of Chicago
Abstract: Large-scale seasonal extreme climate events in one or more important agricultural region can cause major short-term shocks to the global food system. In this talk, we discuss the state of knowledge on the impacts of climate change and extreme events on productivity and food production. We consider plausible global scenarios for 1-in-200-year food-system shocks based on modeling and historical analysis and explore the implications of changing climate on the severity, frequency, and multi-region coincidence of large-scale events in the near future.
As an example of an extreme future event in a major global breadbasket, we discuss recent work looking at the implications of a “new US Dust Bowl”. The 1930s Dust Bowl was the driest and hottest for agriculture in modern U.S. history. Improvements in farming practices have dramatically increased crop productivity, but yields today are still tightly linked to climate variation and the impacts of a 1930s drought on current and future agriculture remains unclear. Simulations suggest that Dust-Bowl-type droughts today would have unprecedented consequences for agricultural productivity, with single-year losses 50% larger than the severe drought of 2012. Damages at these extremes are highly sensitive to temperature, worsening by 25% with each degree C of additional warming. Warmer temperatures with even average precipitation lead to maize losses equivalent to 1936 (the worst season for agriculture in >100 years), and warmer temperatures with consecutive droughts could make up to 85% of rainfed maize at risk to changes that may persist for decades.
Session 1.1: Oral Presentation
Title: Dynamic seasonal climate forecast driven probabilistic maize yield prediction over East Africa
Author: Geoffrey Ogutu et al
1 Earth System Sciences (ESS) Group, Wageningen University, Netherlands, geoffrey.ogutu@wur.nl, wietse.franssen@wur.nl, iwan.supit@wur.nl, ronald.hutjes@wur.nl, 2 Kenya Meteorological Service, Kenya
Abstract: Assuming that skill in seasonal climate prediction would translate to skill in impacts prediction, we use ECMWF system-4 ensemble seasonal climate hindcasts for the period 1981 to 2010 at different initialization dates before sowing to generate 15-ensemble yield predictions using a crop model (WOFOST) implemented for water-limited production and single season simulation. Yield predictions are validated against reference simulations from the same model forced by the WATCH Forcing Data ERA-Interim (WFDEI) at each grid point focussing on the dominant sowing dates in the northern region (July), equatorial region (March-April) and in the southern region (November-December). We use mean deviation, relative bias (pbias), Ranked Probability Skill Score (RPSS) and Relative Operating Curve skill Score (ROCSS) to assess regions of useful probabilistic prediction, mean errors, probabilistic skill, and tercile forecast skill respectively.
Mean yield deviations ≥500 Kg ha-1 in the equatorial regions and coastal Kenya indicate areas of beneficial probabilistic prediction. High deviation may also result from double cropping seasons corresponding to the bi-modal rainfall regimes. Under-prediction exists in all regions. Regions of high deviation show highest pbias (-20% to -80%). RPSS are low and significant in few grid points at lead month-0 except over Ethiopian highlands. Only November harvests over Ethiopia and August harvests in the equatorial regions possess significant above and below normal skill at lead month-0 and 1. From the sample sowing and harvest date combinations considered, we conclude that despite short lead times of useful pre-sowing skill, there is potential to improve yield predictions.
Session 1.1: Oral Presentation
Title : Interdisciplinary and cross-scales Agroclimatic assessment across the U.S. Corn Belt: What have we learnt?
Authors : Xing Liu1 and D. Niyogi2
1 Department of Agronomy, Purdue University, USA, 2Department of Agronomy and Department of Earth, Atmospheric and Planetary Sciences, Purdue University, USA
Abstract: The U.S. Corn Belt produces nearly a third of the global corn supply and contributes nearly $150 billion annually to the U.S economy. Our group has been actively leading efforts for: (i) the synthesis related to understanding the role of weather and climate on the corn yield, and (ii) developing tools and products for making this understanding usable in the context of improved predictions by the meteorologists and improved utilization of the information by crop producers. This presentation will provide an overview of these efforts related to crop modeling and regional agro-climatic analysis for the U.S. Corn Belt, as well as developing tools and delivering datasets for farmers and researchers. The presentation will also focus on the findings from a interdisciplinary project (Making climate information Useful to Useable-U2U), we co-lead as a collaboration with crop modelers, agronomists, atmospheric and social scientists, economists, and communication / extension experts. Studies have been underway related to: crop modeling at field scale and regi nal scale for both contemporary and future time periods, Agro-climatic dataset development, and integrating crop simulation into land surface models for applications within weather and regional climate simulations, including local and regional drought assessment and impact on global trades. The presentation will share our findings, our experience from these multiyear interdisciplinary and cross-scales agroclimatic studies. We will also discuss the limitations, uncertainties, and our perspective on future in the crop modeling studies.
Session 1.1: Oral Presentation
Title: Forecasting effects of weather extremes: El Nino’s influence maize yields in Mexico
Authors: Gideon Kruseman1, K Sonder1, and V. M. Hernández Rodríguez 1
1 CIMMYT
Abstract: Southern Africa is facing a second year of drought in succession. Governments and private traders in the region are looking ahead where to purchase white maize to cover expected food shortages. Mexico is one of the most important white maize producing countries and in normal years has a surplus to trade on the world market. However, the severity of the 2015-2016 El Nino is likely impact on precipitation and temperature in the major rain-fed maize production areas, causing drought and/or heat stress. By forecasting the spatially explicit yield effects, policies can be put in place to ensure that production levels are sufficient to feed Mexico and have a surplus to alleviate the food shortages in Africa.
We use a combination of GIS tools, econometric techniques and crop-growth models to forecast yields and production levels.
Session 1.1: Oral Presentation
Title : Integrated assessment of drought and adaptation scenario impacts on crop production in Austria
Authors: Hermine Mitter1, E. Schmid1, and U. A. Schneider2
1 Institute for Sustainable Economic Development, Department of Economics and Social Sciences, University of Natural Resources and Life Sciences Vienna, Austria, 2 Center for Earth System Research and Sustainability, Research Unit Sustainability and Global Change, University of Hamburg, Germany
Abstract: Drought information systems for agricultural decision-making typically focus on agronomic indicators. We extend these approaches by providing an integrated drought assessment framework covering Austrian cropland at a 1 km grid resolution to (i) quantify the impacts of three drought scenarios in the period 2010-2040 on crop production in Austria, (ii) identify optimal crop production portfolios for drought adaptation considering farmers’ risk aversion, and (iii) calculate the economic value of drought information. The assessment framework consists of a statistical climate model, the bio-physical process model EPIC, a crop gross margin calculation script, a portfolio optimization model, and the computation of the economic value of drought information. At national level, average annual dry matter crop yields in optimal crop production portfolios range between 7.6 and 8.0 t/ha, depending on the drought scenario and the level of risk aversion. Average crop gross margins are between 403 and 473 €/ha. Moderate fertilization intensity is the most frequently chosen management practice in the crop production portfolios, regardless of the drought scenario and the risk aversion level. Average annual values of drought information increase with severity of drought scenario and risk aversion but differ by crop production region. The highest values of drought information (above 200 €/ha) are concentrated in the semi-arid eastern parts of Austria and reveal the opportunity costs of lacking this information for crop management choices. The investigations on potential impacts and effective drought adaptation measures may intensify farmers’ adaptation efforts and inform water resources management and policies.
Session 1.1: Oral Presentation
Title: Multiple crop model ensembles for improving broad-scale yield prediction with Bayesian model averaging
Authors: Xiao Huang1, G. Huang1, C. Yu1, and X. Li1
1 Center for Earth System Science, Key Lab of Earth System Numerical Simulation, Tsinghua University, China
Abstract: Process-based crop models are popular tools to explore the impact of climate change and crop management on crop growth. However, accurate simulation of crop production from single crop model remains challenging over large geographic regions due to different sources of uncertainty. We present the method of Bayesian model average (BMA) for multiple crop-growth model ensembles to provide more reliable predictions of maize yields in Liaoning Province, China, where about 2,200,000 hectares (ha) maize are planted in the 148,000 km2 territory. We apply photosynthesis-oriented WOFOST model, water-orientd AquaCrop model and nitrogen-oriented DNDC model to independently generate the ensemble predictions of the county-level maize yields. The integrated probabilistic prediction is therefore achieved by the linear combination of the three ensembles using the BMA weights. This integrating approach results in an improved accuracy and precision than any single model over the whole region, which demonstrates that the BMA framework effectively compensates the uncertainty of single model simulation and takes advantage of each competing model for reliable prediction. Furthermore, the rationality of the set of BMA weight values is evaluated in comparison with regional precipitation, fertilization and sunshine duration data. We find these values suit well with the regional limiting factor, e.g. AquaCrop model generally obtain high weight value in counties with frequent droughts, while WOFOST is the dominated ensemble in areas with radiation deficit. Compared with simple average method, the results show that the BMA framework is powerful in computing ensemble weights and explaining the mechanism beyond the observed data.
1. Poster Presentation: Session 1.1
Title: Modelling frost damage in major sugar beet growing area of Khorasan province, Iran
Authors: Reza Deihimfard1 and S.R. Moghaddam1
1 Shahid Beheshti University, Iran
Abstract: A simulation study was performed at 8 sites of Khorasan province, in the northeast of Iran where frost damage is one major challenge for a successful cultivation of autumn-winter sugar beet. Accordingly, a modified and validated version of SUCROS model was used to estimate potential yield of sugar beet and frost events over a historical period from 1993 to 2009 under four autumn sowing dates (DOY274, DOY295, DOY314, DOY334) and two spring sowing dates (DOY64 and DOY124). Results indicated that frost damage in autumn sowing dates ranged from 62.5 to 100% at Neyshabour and Ghochan, respectively. According to the cumulative probability distribution of the number of frost events, the highest probability of no frost event obtained at Neyshabor (0.56) and Mashhad (0.5) and the lowest obtained at Bojnord (0.06) and Ghochan (0.0). There was a large variability among sites and years in terms of frost intensity and duration. Minimum temperate reached to -25°C in some years at Mahshad. In contrast, longest frost event occurred at Torbat Jam in 2005 for 39 consecutive days. Although autumn sowing dates showed better performance than spring sowing dates in terms of fresh storage organ yield (~109.9 t ha-1 vs. ~78.4 t ha-1), however, the risk of frost stress under autumn and winter sowing dates are quite high at the all study sites. Accordingly, it is recommended the farmers select an optimum sowing window between February and March during which much lower frost events would be occurred.
2. Poster Presentation: Session 1.1
Title: Impact of extreme events on Russian agriculture
Authors: Paraskevi Giannakaki1 and P. Calanca1
1 Agroscope, Institute for Sustainability Sciences, Climate and Air Pollution Control Group, Switzerland
Abstract: Agriculture in Russia is growing fast and the country has become a major grain exporter. This study is focused on the southwestern part of the Russian Federation (Southern, Central and Volga Districts) where the main grain production occurs (wheat, maize and barley). Reliable forecasts of grain yields in the Russian Federation are therefore important in the context of both world food market and food security. For this, the ability to correctly take into account the effects of extreme events is essential, because grain production in Russia has proved to be highly vulnerable to droughts and heat waves over the past years. Moreover, due to climate change the risk of heatwaves is increasing in the area which implies a decrease in grain production.
In this contribution we examine possibilities to monitor the effects of extreme events by means of agroclimatic indicators, which could serve for improving statistical models as well as for the post-processing of dynamical crop model outputs. Daily climate data from 86 weather stations were used to calculate indices of climate extremes under climatic conditions of the recent past (1980-2014). Moreover, a trend analysis of a set of agroclimatic indices for field crop production is performed. This study is part of a new ERA.NET RUS Plus project “Impact of extreme events and climate change on Russian agriculture, economic implication and adaptions”.
3. Poster Presentation: Session 1.1
Title: Regional assessment of observed, simulated and projected climate over South Punjab
Authors: Ghulam Rasul1 , B. Ahmad1 , A. Ahmad2 , T. Khaliq2 , and G. Hoogenboom3
1 Pakistan Meteorological Department, Pakistan, 2 University of Agriculture, Faisalabad, Pakistan, 3 Institute for Sustainable Food Systems, University of Florida, USA
Abstract: Extraction of regional climate information from Global Circulation Models (GCMs) require robust selection and integration. Following AgMIP phase II protocols, background daily climate time series (1980-2010) from the AgMERRA datasets are obtained to fill in missing/flagged observation data. Future climate scenarios under two RCPs (4.5 and 8.5) are derived from the latest 29 IPCC climate models and downscaled for use in the target regions. Subsetting of GCMs from 29 to 5 is done based on scatter of mean temperature change and mean precipitation change for the entire growing and harvesting season of cotton-wheat system in South Punjab. Models bearing precipitation projection of not more than 200% are selected. In addition, replication of monsoon in historical climate is also taken into consideration while selecting GCMs. Downscaling of regional climate is done by shifting mean and variability scenarios using the stretched distribution approach that is related to quantile mapping. Post selection and downscaling results under RCP8.5 and CCSM4 model (with cool/dry characteristics) suggest a 2.5°C rise in the maximum and minimum temperatures with 9% decrease in the precipitation amount in the 2040-2069 projected period over the region. The projected increase in temperature and the corresponding decrease in the precipitation regime give clues regarding devastation in the agricultural yield in the 2040-2069 projection period over the region.
Keywords: South Punjab, GCMs, RCPs, AgMERRA, Downscaling, Projections.