Early prediction of wheat yield using seasonal weather forecasts and the static stress drought index

Early prediction of wheat yield using seasonal weather forecasts and the static stress drought index | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Early prediction of wheat yield using seasonal weather forecasts and the static stress drought index Vanesa García-Gamero, Andrés Peñuela, Adolfo Peña, Tom Vanwalleghem This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-2742457/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Crop yield prediction considering soil moisture as a proxy for water supply remains crucial for global food security. This study evaluates the potential of using seasonal weather forecasts combined with a drought index, Static Stress, based on both precipitation and soil moisture conditions to predict winter wheat yield 7 to 1 month in advance in Córdoba (South Spain). First, using observed climate and crop yield data we evaluate the use of Static Stress, as a potential crop yield predictor and compare it to a more traditionally used index, the SPEI, which is only based on precipitation conditions. Then we evaluate the performance of simple linear regression models to predict crop yields from forecasted Static Stress values calculated using weather forecast data from the ECMWF seasonal forecasting system (SEAS5). We find that Static Stress is better correlated to crop yield than SPEI and that Static Stress derived from seasonal forecasts has a good performance (R 2 > 0.5; p-value < 0.05) for crop yield predictions of 4 or fewer months before harvest, i.e., from March to July. In this case study, these results indicate that drought indicators that consider soil moisture conditions are better predictors of crop yields than indicators that only consider precipitation. Furthermore, this study demonstrates the potential of using simple regression models together with mid-term forecasts of the Static Stress index to maximize cereal yields and mitigate drought impacts. Earth and environmental sciences/Hydrology Earth and environmental sciences/Climate sciences/Climate change Earth and environmental sciences/Climate sciences/Hydrology Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Introduction The ongoing armed conflicts and the climate crisis jeopardize global food security 1 . While conflicts are currently viewed as the main driver of food insecurity 2 . Extreme climatic events, such as droughts or flooding, are particularly important however as they significantly impact agricultural production 3 , 4 , 5 at a global scale, and increase the risk of armed conflict 6 , 7 . Weather extremes, and in particular droughts, are expected to become more frequent and severe over the next years due to climate change 8 . In Europe, drought-related crop production losses tripled between 1964–1990 and 1991–2015: from 2.2–7.3% 9 . A study by Jenkins 10 showed that in 2003–2050 annual drought losses will increase by 32% and 57% for short-term and long-term droughts, respectively, in Spain. Spain is a particularly sensitive country to droughts that has been subject to very severe drought events in the last century 11 , 12 causing important agricultural losses 13 . Therefore, there is an urgent need to develop tools to anticipate drought impacts and to support the definition of mitigation measures. Drought is an extreme hydroclimatic event whose definition is not unique 14 . The meteorological definition is the most prevalent, ”a period of more than some particular number of days with precipitation less than some specified small amount” 15 . However, agricultural drought is not only related to rain scarcity but also poor soil and soil moisture conditions that result in adverse crop responses, such as reduced crop yield 16 , 17 . Indeed, deep, fertile soils with a large soil water holding capacity constitute a better buffer against droughts than shallow, eroded soils. Further, drought can also deteriorate the soil conditions by increasing soil erodibility and hence soil loss 18 . The use of drought indices is the most common approach for quantifying and monitoring drought intensity and impacts on agriculture. Most drought indicators are based on precipitation measurements 19 . Among them, the SPI 20 and Standardized Precipitation Evapotranspiration Index (SPEI) 21 are widely used and recommended by the World Meteorological Organization (WMO). However, agricultural droughts, i.e., when crops become affected, usually occur when there is a soil moisture deficit. Jiménez-Donaire et al. 22 evaluated the potential of two novel indices, Static Stress, and Dynamic Stress, that consider soil moisture dynamics and found them better correlated with wheat yield than SPI in Southwestern Spain. Other type of drought indicators are the Combined Drought Indicators (CDI) which encompass meteorological, soil moisture and vegetation condition satellite-derived data and are more adequate to monitor agricultural droughts 23 . A CDI designed by Sepulcre-Canto et al. 23 distinguishes categorical drought warnings based on satellite-derived data. However, the validation of these categorical classes that represent the severity of the impact of drought, such as crop yield losses, is lacking 24 . The main limitations of CDI are that they need satellite products of high resolution and, especially, that these products are not adequate for future forecasts, and are limited to historical observations to make future predictions. Numerous studies have linked drought indices to crop yield variability in Spain. García-León et al. 25 concluded that remote sensing data related to vegetation health provided more accurate information on local drought conditions compared to drought indices based solely on rainfall, largely because the former incorporates local biophysical and climate factors inherently. Peña-Gallardo et al. 26 assessed the impact of drought also on barley and wheat, two rainfed crops in Spain. They found a good correlation between crop yield and SPI and SPEI, however, the weakest correlations were found in South Spain. While these studies show good results in terms of correlating crop yields with drought indices, these results are based on past observed data. To support land management and in particular, the definition of measures to reduce the impact of droughts in agriculture, we should further explore the potential of estimating drought indices from mid-term weather forecasts, i.e. several months in advance. Forecasting droughts that occur during the most sensitive stages of crops is paramount for soil and land management. In this sense, seasonal forecasts can act as a support for stakeholders and policymakers in making more optimal decisions several months in advance. Seasonal forecasts can not only be used to develop drought early warning systems but also for drought impact forecasts 27 , such as crop yield losses. In general, in the literature there is a lack of studies, evaluating the potential of making future predictions of crop yields using drought indices derived from mid-term (or seasonal) weather forecasts. In this study, we present the first study to evaluate the potential of using seasonal weather forecasts combined with a drought index to predict cereal crop yields. Specific objectives are (i) to evaluate the correlation between the Static Stress, based on both precipitation and soil moisture conditions, and compare it against the more traditional indicator SPEI only based on precipitation and (ii) to develop and evaluate a regression model to predict winter wheat yield for a lead time from 0 to 7 months in Córdoba (South Spain). The model is based on the use of the Static Stress and seasonal weather forecasts data provided by the ECMWF SEAS5. The information provided here and the proposed methodoly may benefit for farmers and stakeholders to maximize cereal yields and mitigate drought impacts. Material And Methods A schematic overview of the applied methodology is shown in the flowchart in Fig. 1 and explained in detail below. The top panel (A) provides a framework for evaluating drought indexes. In this sense, two datasets of observed data were used in this study: observed winter wheat yield (Fig. 1 box 1), and observed weather data (Fig. 1 box 2). Observed weather data were used to estimate soil moisture using a soil moisture model (Fig. 1 box 3). The drought indexes that are estimated using observed weather data are the SPEI, for two different periods, and the Statis Stress Index using soil moisture (Fig. 1 box 4). Model performance is measured by the coefficient of determination (R 2 ) to evaluate which drought index is better correlated to the crop yield (Fig. 1 box 5). Then, the lower panel (B) provides a framework to define eight linear regression models (one per forecast lead time) using the observed weather data dataset combined with seasonal forecasts (Fig. 1 box 6) to predict annual wheat yield. Observed values (until the date of issue of the forecast) are merged with seasonal forecasts from 7 months lead time (only seasonal forecast data) to 0 months lead time (only observed data) to cover the winter wheat growing season (December to July) (Fig. 1 box 7). These 7-month weather data are used to simulate soil moisture (Fig. 1 . Box 8) and calculate the drought index (Fig. 1 . Box 9). To evaluate the relationship between observed winter wheat yield and the drought index, calculated as the mean value obtained for the set of 25 members or forecast provided by the ECMWF at each forecast lead time, we computed R 2 (Fig. 1 box 10). R 2 values are obtained for the 8 lead times that we define previously. 1.1. Region of study The area of study is the province of Córdoba, located in South Spain. It was selected because of its large land extension dominated by rainfed crops, in particular, herbaceous crops such as winter wheat (Fig. 2 ). The climate of this region is Mediterranean, which is characterized by hot and dry summers and cold winters. The average annual temperature is 18°C 22 , with an average annual precipitation of 602 mm, with a standard deviation of 184 mm (1961–2018). Figure 3 A shows the time series of precipitation during the hydrological year of the study period 2002–2020 and Fig. 3 B the corresponding precipitation for different stages of wheat development: from the end of anthesis (March) to grain filling (April-June). The average precipitation during the selected period was 605 mm, ranging from 314 mm to 1062 mm. The standard deviation was 192 mm. 1.1 . Datasets 1.1.1. Observational crop yield dataset For the 2002–2020 study period winter wheat yield was compiled (Fig. 1 box 1). The average winter wheat yield was 2657 kg ha − 1 28 in the province of Córdoba (dashed line in Fig. 3 C). Wheat in this area is usually planted from late November to early December and is usually harvested in late June to early July. In this study, the growing season considered was from December 1st to July 1st . In Spain, Andalusia accounts for 67.7% of winter wheat production 29 . As shown in Fig. 3 C, winter wheat yields have remained relatively stable since the beginning of the study period. However, drops in yields in specific years, such as that of 2005 can be observed. As reported by the Spanish State Meteorological Agency (AEMET), 2005 was the driest year in the study period (1961–2018) 30 , with a mean annual precipitation of 451 mm for peninsular Spain, which represents 70% of the average (640 mm). This severe precipitation deficit together with another drought of unprecedented intensity in the 2007–2008 hydrological year 31 caused the onset of a long-lasting drought based on SPI-36 from December 2005 to December 2009 30 . It is remarkable that since 2005, the surface area devoted to winter wheat has declined from 140,000 to 40,000 ha (Fig. 3 D). Espinosa-Tasón et al. 32 studied the economic impact of multiyear droughts (2005–2008) in Andalusia. The area dedicated to cereals has decreased since 2005, as well as rainfed and irrigated agricultural production. Nonetheless, some irrigated crops experienced noticeable increases in production in 2005–2008 due to agricultural expansion since 2000 32 . 1.1.2. Observational weather dataset The daily weather data (Precipitation, P, and Reference evapotranspiration, ET 0 ) during 2002–2020 was obtained from Córdoba station (37.86° N,4.80°W) (Fig. 1 box 2), at 94 meters above the sea level, of the Agroclimatic Information Network of Andalusia (RIA, https://www.juntadeandalucia.es/agriculturaypesca/ifapa/riaweb/web/estacion/14/6 ). We used the values proposed by Vanderlinden 33 calculated from the soil map of Andalusia for soil moisture content at wilting point (s w ) and field capacity (s fc ). 1.1.3. Forecast dataset The seasonal weather forecast data were obtained from the ECMWF SEAS5 (Fig. 1 box 6). ECMWF provides both real-time seasonal forecasts and historical re-forecasts, i.e. hindcasts, generated with the last version of the system, i.e. SEAS5, currently used for the real-time forecasts. These two products have a 7-month lead time starting the 1st day of the month. They have a spatial resolution of 1⁰ x 1⁰ (36 km approximately) and are composed of 25 members for the hindcast and 50 for the forecast. In this study, daily precipitation and evapotranspiration are retrieved from 2002 until 2016 (hindcasts) and from 2017 to 2020 (forecasts). For the remainder of the study, we will use only the term forecast to refer to both hindcast and forecast. 1.2. Drought indices 1.2.1. Standardized Precipitation Evaporation Index The Standardized Precipitation Evaporation Index (SPEI-n) (Fig. 1 box 4) is a simple multiscale drought index based on a combination of precipitation and temperature data. SPEI is calculated using a fixed-length moving window based on a monthly difference between precipitation and calculated potential evapotranspiration, using the Hargreaves method based on temperature. In this study, we focus on shorter accumulation periods, or SPEI-3, which is more closely related to the impact of drought on seasonal herbaceous systems 25 . According to the Spanish annual crop calendar 34 , the growing season considered for winter wheat was between December and July. SPEI-3 was quarterly aggregated to capture two different growth stages of the plant: mid or anthesis (January-March) and late or grain-filling stages (April-June) as recommended by García-León et al. 25 . SPEI values range from − 3 (extremely dry) to 3 (extremely wet) 25 . The SPEI values were obtained using the R package SPEI ( https://cran.r-project.org/web/ packages/SPEI/index.html). 1.2.2. Static Stress The Static Stress (Fig. 1 box 4), ζ , estimates the plant water stress related to soil moisture conditions 22 . It is determined as Eq. (1) 17 : $$\genfrac{}{}{0pt}{}{\zeta =0 for s>s*,}{\zeta =1 for s<{s}_{w}}$$ 1 where the ζ is zero when soil moisture, s, is above the incipient stomatal closure, s*, and is equal to one when soil moisture is below the permanent wilting point, s w . For soil moisture values between s* and s w , the general expression is as Eq. ( 2 ): $$\zeta \left(t\right)={\left[\frac{{s}^{*}-s\left(t\right)}{{s}^{*}-{s}_{w}}\right]}^{q}, for {s}_{w}\le s\left(t\right)\le {s}^{*},$$ 2 where q is a measure of the non-linearity of the effects of soil moisture deficit on plant conditions. In this study, we use a value of 1 which implies a linear relation between static water stress and soil moisture deficit 22 . Soil moisture dynamics are calculated following the same approach presented by Jiménez-Donaire et al. 22 which consists of a soil water-balance “bucket” model considering infiltration, evapotranspiration, and deep seepage over the root zone (Fig. 1 box 3 and box 8). 1.3. Regression model for wheat yield For each forecast lead time, we used a linear regression model to estimate the crop yield at the end of the growing season. For this purpose, we define 8 scenarios using both observed and forecast data ( Fig. 1 box 6 and 7 and Fig. 4 top to bottom panels). We evaluated the performance of the model to predict the crop yield by the end of the harvest period (1 July) for 7 forecast lead times: from 7 months (starting 1 December) to 1 month (starting 1 June). When the starting month is different from December, e.g., February, observed precipitation and evapotranspiration observed values are used until the time of forecast initialization, e.g. 1 February. Then, for each lead time, we calculated the mean of the selected drought index for the 25 forecast ensemble members (Fig. 4 ). The process is repeated each month until estimations are obtained for a lead time of 0 months, i.e., only observed data (bottom panel in Fig. 4 ). Since we are interested in predicting drought events, no bias correction has been applied to the seasonal weather forecast. The justification for this is that bias correction pushes forecasts to be more like average past weather, therefore bias correction may reduce the “good signal” that may be present in the original forecast in months that will indeed be significantly drier than the average 35 . The correlation is calculated between the forecasted yearly mean value of the drought index and the observed winter wheat yields for lead times between 0 and 7 months. For each lead time, the R 2 value will be presented in the form of a bar plot. Results And Discussion 2.1. Evaluation of Static Stress for Prediction of Wheat Yield Observed Static Stress and SPEI-3 were evaluated and compared based on their capacity to explain winter wheat yield variability (Fig. 1 box 5). The results of the three statistical models with the winter wheat yield data and these drought indices are presented in Fig. 5 . Static Stress shows a good correlation with yield, with an R 2 of 0.57. By comparison, SPEI-3 performs poorly, with an R 2 value of 0.25 and 0.12 for the mid and late stages respectively. It should be noted that 2018 (Fig. 5 . in red) was identified as an outlier and removed from the analysis. Based on the findings of this study, losses in wheat yield can be well explained by the Static Stress for the study period. While this agrees with the results of other studies 22 other factors may also influence yield and hence, the occurrence of a drought event may not necessarily lead to a reduction in yield 36 . This might explain why in 2018 the high crop yield does not correspond to a low Static stress value. Nevertheless, exceptionally intense droughts can considerably increase the probability of yield loss for wheat 36 . Another important difference between Static Stress and SPEI is that the former only indicates the presence of drought stress. Therefore, wet years are characterized by 0 values. In contrast, the relationship between yield and SPEI is overall positive, but especially for Jan-Mar, the relation is more complex, and first positive up till 0 and then negative. This can be explained because dryer years are characterized by more negative values, but wetter than average years can also see reductions in crop yield which explains the decline of yields for values > 0. The highest yields are generally observed for years with precipitation around average, i.e. SPEI-3(Jan-Mar) = 0. For SEPI-3(Apr-Jun), similar behavior can be observed but the trend is less clear. Overall, the results of this study case show that Static Stress is a better predictor of crop yield than SPEI and indicate that drought indicators that consider both precipitation and soil moisture conditions are better predictors of crop yields than indicators that only consider precipitation. 2.2. Forecast of wheat yield Model performance measured by R 2 value per lead-time, based on the mean Static stress calculated using the set of 25 members of the seasonal weather forecasts, and observed past weather data, to predict winter wheat yield is assessed (Fig. 6 ) for each forecast lead time. As expected, R 2 values gradually decrease with longer forecast lead times. In general, the proposed linear regression models show good performance (R 2 > 0.5; p-value < 0.05) for lead times equal to or less than 4 months, i.e. from March to July. Figure 6 shows that the best correlation is obtained for a forecast lead time of 2 months (R 2 = 0.64; p-value < 0.05), i.e. combination of 5 months of observed weather data (December-April) and 2 months of seasonal weather forecast data (May-June). The results for a 3 months lead time (R 2 = 0.58) are similar to those obtained for a 0-month lead time, i.e. using only observed data (R 2 = 0.57). There is a significant difference between the results obtained for 5 months (R 2 = 0.377), i.e. from February to July, and 6 months (R 2 = 0.17), i.e. from January to July. Surprisingly, the correlation is higher for a 2-month lead time than for a 1-month lead time and even when using only observed data (distinguished in Fig. 6 in dark blue; 0-month lead time). This study proposes a simple method to predict yields through a regression method with relatively low data and computational requirements. In the literature, there is a wide range of models used for yield forecasting but regression models showed better results in 74% of studies assessed by Schauberger et al. 37 . Meroni et al. 38 also claimed that regression models outperformed 60% of tested machine learning models in predicting barley, soft, and durum wheat yields in Algeria. Moreover, machine-learning models require larger datasets 38 . To our knowledge, we present the first study that assesses the feasibility of combining drought indices with seasonal forecasts to predict crop yields. A study by Bento et al. 39 used linear regression models and seasonal precipitation and temperature forecasts, instead of drought indices, from SEAS5 to predict wheat and barley yield in the Iberian Peninsula. Similarly to this study, winter wheat yield forecasts only started to gain skill from April onwards, i.e. three months before the harvest. They remark that some variables chosen by their selected regression models are bound to change in the coming decades due to unavoidable warning scenarios and need to be parameterized locally. In Europe, Ceglar and Toreti 40 tested the predictability of drought events during winter wheat production however they did not relate drought indices with crop yield. In their study, skillful and reliable predictions of drought events using SEAS5 forecast with indicators based on SPEI can be made already at the end of winter. In Australia, Jin et al. 41 found significant yield forecast skill improvement at the beginning of the winter wheat cropping season, such as 1st May or 1st June by using the seasonal climate forecast model ACCESS-S1 and a process-based crop model. Operational yield forecasting systems exist at a regional scale, such as the European Union (Mars 42 ), assessing the increasingly variable climatic conditions that are impacting crop production systems throughout the EU, and on a global scale 43 . Despite the challenges associated with implementing a global forecasting system, accurate information could be beneficial to agribusinesses and food security in regions where crop yields are susceptible to climate change and food markets are not connected to international trade 43 . Next-generation crop yield forecasting systems are increasingly sophisticated and include machine learning techniques featuring neural networks 37 . Khaki and Wang 44 used deep neural networks for forecasting crop yield pointing out the sensitivity of the model to weather data origin. They also highlighted that even when their model outperformed other models, it is more accurate than explanatory. In Spain, efforts need to be made to make such information available to decision-makers in a form that is understandable and useable. The use of in-season decision support systems that incorporate crop models might improve the link between agricultural practices and climate information 45 . In several countries, to reduce the risk of crop failure and increase farmers' profits, seasonal climate forecasts are linked to crop simulation models to provide potentially useful information for stakeholders. Since 2019, the commercial platform CropProphet used ECMWF forecast data for their crop yield and production forecast models, in the USA. Their resulting 15-day yield forecast models provide the advantage of being ahead of market expectations and futures market trading 46 . Furthermore, another platform, Planet, supplies daily corn yield forecasts of the growing season also in the USA. This is done by using an algorithm that makes optimal use of archival records, giving their statistical models an unparalleled trove of data to correlate with historical yields 47 . In Australia, Wheatcast™ integrated analytics systems lead to a yield forecast at any time and scale 48 . While this study show the potential for the practical application of the proposed method some limitations of the results must be considered. Winter wheat yields correspond to the average of the whole province of Cordoba, not to specific areas. With higher spatial resolution yield data we could have better link changes in yield due to local changes in weather and soil moisture. Because of this, we also used the same soil and plant parameters in the soil water balance model and to calculate plant stress were unique. Conclusions In this study, we assessed the potential of using seasonal weather forecasts and the Static Stress drought index to predict winter wheat yield in South Spain. First, we find that Static Stress is better correlated to crop yield than SPEI and that Static Stress derived from seasonal forecasts shows a good performance in predicting crop yields for 4 or fewer months before harvest, i.e., from March to July. These results indicate that drought indicators that consider soil moisture conditions are better predictors of crop yields than indicators that only consider precipitation. Furthermore, this study demonstrates the potential of using the Static Stress index derived from seasonal forecasts to predict crop yields. Knowing the yield before harvest can help stakeholders maximize production while avoiding expenses that might affect small farmers and economies. Declarations Data availability Statement . The datasets generated during and/or analyzed during the current study are available from the corresponding author on request. Acknowledgments This work was supported by the Spanish Ministry of Science, Innovation, and Universities co-financed by the EU through the project “Quantifying the impact of soil erosion and climate change on soil security by using alternative fallout radionuclides” (PID2019-109924RB-I00/AEI/10.13039/501100011033). García-Gamero, Peñuela, and Vanwalleghem acknowledge financial support from the Department of Agronomy, of the Spanish Ministry of Science and Innovation, the Spanish State Research Agency, through the Severo Ochoa and María de Maeztu Program for Centers and Units of Excellence in R&D (Ref. CEX2019-000968-M). The authors wish to thank the Copernicus Climate Change and Atmosphere Monitoring Services for providing the seasonal forecasts generated by the ECMWF seasonal forecasting systems (SEAS5). Neither the European Commission nor ECMWF is responsible for any use that may be made of the Copernicus information or data it contains. Author contributions VGG and TV designed the research. VGG and 1 AP review the literature and collected the data. VGG, 1 AP, and TV designed the analytical strategy. VGG conducted the data analyses and produced the results. VGG performed the interpretation of the results with the contribution of 1 AP, 2 AP and TV. VGG wrote a first draft of the manuscript and thereafter 1 AP, 2 AP , and TV made subsequent revisions to the original text and figures. Additional information Competing interests The authors declare no competing interests. References Scheffran, J. & Battaglini, A. Climate and conflicts: the security risks of global warming. Reg Env. Chang. (2011). doi: 10.1007/s10113-010-0175-8 Food Security Information Network. GLOBAL REPORT ON FOOD CRISES . (2022). Zhu, P. et al. Warming reduces global agricultural production by decreasing cropping frequency and yields. Nat. Clim. Chang. 12, (2022). Dasgupta, S. & Robinson, E. J. Z. Attributing changes in food insecurity to a changing climate. Sci. Rep. 1–11 (2022). doi: 10.1038/s41598-022-08696-x Krishnamurthy R, P. K., Fisher, J. B., Choularton, R. J. & Kareiva, P. M. Anticipating drought-related food security changes. Nat. Sustain. (2022). doi: 10.1038/s41893-022-00962-0 Ge, Q. et al. Modelling armed conflict risk under climate change with machine learning and time-series data. Nat. Commun. 13, 1–8 (2022). Von Uexkull, N., Croicu, M., Fjelde, H. & Buhaug, H. Civil conflict sensitivity to growing-season drought. Proc. Natl. Acad. Sci. U. S. A. 113, 12391–12396 (2016). Hasegawa, T. et al. Extreme climate events increase risk of global food insecurity and adaptation needs. Nat. Food 2, 587–595 (2021). Armada Brás, T., Seixas, J., Carvalhais, N. & Jägermeyr, J. Severity of drought and heatwave crop losses tripled over the last five decades in Europe. Environ. Res. Lett. 16, 65012 (2021). Jenkins, K. Indirect economic losses of drought under future projections of climate change: A case study for Spain. Nat. Hazards 69, 1967–1986 (2013). González-Hidalgo, J. C. et al. High-resolution spatio-temporal analyses of drought episodes in the western Mediterranean basin (Spanish mainland, Iberian Peninsula). Acta Geophys. 66, 381–392 (2018). Vicente-Serrano, S. M. Spatial and temporal analysis of droughts in the Iberian Peninsula (1910–2000). Hydrol. Sci. J. 51, 83–97 (2006). UPA. Agricultura familiar en España. Anuario 2017 . (2017). Wilhite, D. A. & Glantz, M. H. Understanding the Drought Phenomenon: The Role of Definitions. Water Int. 10, 111–120 (1985). Great Britain Meteorological Office. Meteorological Glossary . Great Britain Meteorological Office (Chemical Publishing Co., 1951). doi: 10.1002/j.1477-8696.1964.tb02532.x Kang, Y., Khan, S. & Ma, X. Climate change impacts on crop yield, crop water productivity and food security - A review. Prog. Nat. Sci. 19, 1665–1674 (2009). Porporato, A., Laio, F., Ridolfi, L. & Rodriguez-Iturbe, I. Plants in water-controlled ecosystems: active role in hydrologic processes and response to water stress III. Vegetation water stress. Adv. Water Resour. 24, 725–744 (2001). Baumhardt, R. L. & Blanco-Canqui, H. Soil: Conservation Practices. Encycl. Agric. Food Syst. 5, 153–165 (2014). Dutra, E. et al. Global meteorological drought-Part 1: Probabilistic monitoring. Hydrol. Earth Syst. Sci. 18, 2657–2667 (2014). Bhanu Kumar, O. S. R. U., Ramalingeswara Rao, S., Ramaganathan, S. & Raju, S. S. Role of Intra-seasonal Oscillations on Monsoon Floods and Droughts over India. Asia-Pacific J. Atmos. Sci 46, 21–28 (2010). Vicente-Serrano, S. M., Beguería, S. & López-Moreno, J. I. A multiscalar drought index sensitive to global warming: The standardized precipitation evapotranspiration index. J. Clim. 23, 1696–1718 (2010). Jiménez-Donaire, M. del P., Giráldez, J. V. & Vanwalleghem, T. Evaluation of Drought Stress in Cereal through Probabilistic Modelling of Soil Moisture Dynamics. Water 12, 2592 (2020). Sepulcre-Canto, G., Horion, S., Singleton, A., Carrao, H. & Vogt, J. Development of a Combined Drought Indicator to detect agricultural drought in Europe. Nat. Hazards Earth Syst. Sci. 12, 3519–3531 (2012). Zargar, A., Sadiq, R., Naser, B. & Khan, F. I. A review of drought indices. Environ. Rev. 19, 333–349 (2011). García-León, D., Contreras, S. & Hunink, J. Comparison of meteorological and satellite-based drought indices as yield predictors of Spanish cereals. Agric. Water Manag. 213, 388–396 (2019). Peña-gallardo, M., Vicente-serrano, S. M., Domínguez-castro, F. & Beguería, S. The impact of drought on the productivity of two rainfed crops in Spain. Nat. Hazards Earth Syst. Sci. 1215–1234 (2019). Sutanto, S. J., van der Weert, M., Wanders, N., Blauhut, V. & Van Lanen, H. A. J. Moving from drought hazard to impact forecasts. Nat. Commun. 10, 1–7 (2019). Ministerio de Agricultura Pesca y Alimentación. Encuesta sobre Superficies y Rendimientos Cultivos (ESYRCE). Boletín Mensual de Estadística (2022). Available at: https://www.mapa.gob.es/es/estadistica/temas/estadisticas-agrarias/agricultura/esyrce/ . (Accessed: 15th October 2022) Ministerio de Agricultura Pesca y Alimentación. Estadística Anual De Superficies y Producciones de cultivos . (2022). Romero Fresneda, R. et al. Comportamiento de las precipitaciones en España y periodos de sequía (periodo 1961 –2018) . Comportamiento de las precipitaciones en España y periodos de sequía (periodo 1961–2018) (Agencia Estatal de Meteorología, 2020). doi: 10.31978/666-20-006-0 Mestre, A. & Moreno, J. V. El Seguimiento de la Sequía en Eastaciones y Sobre Grandes Cuencas a Través de Uso del índice SPI. Boletín AME 36–41 (2009). Espinosa-Tasón, J., Berbel, J., Gutiérrez-Martín, C. & Musolino, D. A. Socioeconomic impact of 2005–2008 drought in Andalusian agriculture. Sci. Total Environ. 826, (2022). Vanderlinden, K. ANÁLISIS DE PROCESOS HIDROLÓGICOS A DIFERENTES ESCALAS ESPACIO-TEMPORALES. (UNIVERSIDAD DE CÓRDOBA, 2001). Oteros, J., García-Mozo, H., Botey, R., Mestre, A. & Galán, C. Variations in cereal crop phenology in Spain over the last twenty-six years (1986–2012). Clim. Change 130, 545–558 (2015). Peñuela, A., Hutton, C. & Pianosi, F. Assessing the value of seasonal hydrological forecasts for improving water resource management: Insights from a pilot application in the UK. Hydrol. Earth Syst. Sci. 24, 6059–6073 (2020). Leng, G. & Hall, J. Crop yield sensitivity of global major agricultural countries to droughts and the projected changes in the future. Sci. Total Environ. 654, 811–821 (2019). Schauberger, B., Jägermeyr, J. & Gornott, C. A systematic review of local to regional yield forecasting approaches and frequently used data resources. Eur. J. Agron. 120, 126153 (2020). Meroni, M., Waldner, F., Seguini, L., Kerdiles, H. & Rembold, F. Yield forecasting with machine learning and small data: What gains for grains? Agric. For. Meteorol. 308–309, (2021). Bento, V. A. et al. Persistence versus dynamical seasonal forecasts of cereal crop yields. Sci. Reports 2022 121 12, 1–11 (2022). Ceglar, A. & Toreti, A. Seasonal climate forecast can inform the European agricultural sector well in advance of harvesting. Clim. Atmos. Sci. (2021). doi: 10.1038/s41612-021-00198-3 Jin, H., Li, M., Hopwood, G., Hochman, Z. & Bakar, K. S. Improving early-season wheat yield forecasts driven by probabilistic seasonal climate forecasts. Agric. For. Meteorol. 315, 108832 (2022). van der Velde, M. & Nisini, L. Performance of the MARS-crop yield forecasting system for the European Union: Assessing accuracy, in-season, and year-to-year improvements from 1993 to 2015. Agric. Syst. 168, 203–212 (2019). Doi, T., Sakurai, G. & Iizumi, T. Seasonal Predictability of Four Major Crop Yields Worldwide by a Hybrid System of Dynamical Climate Prediction and Eco-Physiological Crop-Growth Simulation. Front. Sustain. Food Syst. 4, (2020). Khaki, S. & Wang, L. Crop Yield Prediction Using Deep Neural Networks. Front. Plant Sci. 1, 621 (2019). Guarin, J. R. & Asseng, S. Wheat crop modelling to improve yields . (Burleigh Dodds Science Publishing Limited, 2017). doi: 10.19103/as.2016.0004.27 Dutton, J. Corn Yield Forecast: The #1 Critical Equation | CropProphet. (2020). Available at: https://www.cropprophet.com/corn-yield-forecast-model/ . (Accessed: 26th October 2022) van der Schalie, R. CornCaster: How Planet’s Yield Forecasting Solution Is Helping Agriculturists And Economists Get Ahead Of This Year’s Harvest. (2022). Available at: https://www.planet.com/pulse/corncaster-how-planets-yield-forecasting-solution-is-helping-agriculturists-and-economists-get-ahead-of-this-years-harvest/ . (Accessed: 29th October 2022) Lawes, R. et al. Graincast™: monitoring crop production across the Australian grainbelt. Crop Pasture Sci. (2022). doi: 10.1071/cp21386 Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-2742457","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":188322384,"identity":"7a1fa289-f1ca-48ae-b338-30811c6fdd46","order_by":0,"name":"Vanesa García-Gamero","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyUlEQVRIiWNgGAWjYDACZuYDYJqfeC3sbQlgWrKBaC08ZwzAtMEBYnXIRyQYv/jxyy5x843khw8YKuoIazG8kZBm2duXnLjtRpqxAcOZw0RomZFwzIC3h9nY7EYOmwRjGxHOM5yR2Gb4t6fe2HgGSMs/Ihwmz3OY+THPj8NyBhIgLQ3MhLUYsLexMcs2HJeTOPPM2CDhGBF+kW/m//zxzZ9qHv52YIh9qCHCYcDoAPkayksgrAFoSwMD8weGP8QoHQWjYBSMghELAFyjO5z3c806AAAAAElFTkSuQmCC","orcid":"","institution":"University of Córdoba","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Vanesa","middleName":"","lastName":"García-Gamero","suffix":""},{"id":188322386,"identity":"de19329b-d45a-436d-8fdc-7182301359f2","order_by":1,"name":"Andrés Peñuela","email":"","orcid":"","institution":"University of Córdoba","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Andrés","middleName":"","lastName":"Peñuela","suffix":""},{"id":188322387,"identity":"4d660bcd-cd7b-4a7a-8e6d-6bed6bcd862a","order_by":2,"name":"Adolfo Peña","email":"","orcid":"","institution":"University of Córdoba","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Adolfo","middleName":"","lastName":"Peña","suffix":""},{"id":188322388,"identity":"93e7de98-7ce2-4d71-b681-7ad0dff3cd28","order_by":3,"name":"Tom Vanwalleghem","email":"","orcid":"","institution":"University of Córdoba","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Tom","middleName":"","lastName":"Vanwalleghem","suffix":""}],"badges":[],"createdAt":"2023-03-27 13:59:19","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-2742457/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-2742457/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":35294582,"identity":"a40beb8c-6731-4451-bc45-f3a6c4629b38","added_by":"auto","created_at":"2023-04-04 22:08:44","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":145749,"visible":true,"origin":"","legend":"\u003cp\u003eFlow chart showing the data and methods used in this study. The top panel (A) indicates the process for evaluating the correlation between the drought indexes and the observed crop yield: (1) Observed crop yield and (2) Observed weather data for the 2002-2020 period were used to calculate, on one hand, (3) using a soil moisture model, (4) the Static Stress index and the other hand, SPEI for two different periods. Then, (5) Model performance measured by R\u003csup\u003e2\u003c/sup\u003e\u0026nbsp; concluded which drought index was selected by comparing the previously mentioned indexes.\u0026nbsp; Then, the lower panel (B) indicates that to predict crop yield (6) Observed weather data are merged with seasonal forecasts to generate (7) 8 different scenarios. Scenarios vary from only observed weather data used (lead time = 0 months) to only seasonal weather forecasts used (lead time = 7 months) to cover the crop growing season. Then, (8) using the soil moisture model, (9) the drought index is calculated. Finally, (10) models performance measured by\u0026nbsp; R\u003csup\u003e2\u003c/sup\u003e was obtained considering the observed crop yield and the calculated mean of the drought index for the crop growing season obtained for the set of\u0026nbsp; 25 members or forecast provided by the ECMWF\u0026nbsp; at each forecast lead time.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-2742457/v1/01f37136c1e2c33e90053a8b.png"},{"id":35294583,"identity":"ee06b4eb-b7fb-4bd3-990f-9ca04077a3ff","added_by":"auto","created_at":"2023-04-04 22:08:44","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":32057,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of herbaceous crops including winter wheat in the province of Córdoba, as derived from the Spanish Land Cover Information System (SIOSE,2014).\u003c/p\u003e","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-2742457/v1/fe1781af3ebc91739165b81e.png"},{"id":35294584,"identity":"5c853b58-4090-47c3-8928-42e1aeaf73bc","added_by":"auto","created_at":"2023-04-04 22:08:44","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":16652,"visible":true,"origin":"","legend":"\u003cp\u003eThe plot shows (A) the observed annual precipitation (hydrological year), (B) the observed monthly precipitation to capture the late mid- or anthesis (March) and late or grain filling (April-June) growth stages of the crop, (C) annual and mean (dashed line) Yield (kg ha\u003csup\u003e-1\u003c/sup\u003e) and (D) Area (ha) of winter wheat during the study period. Winter wheat annual yield (kg ha\u003csup\u003e-1\u003c/sup\u003e) and area (ha) for the selected province were obtained from the Spanish Ministry of Agriculture, Fisheries and Food (available at \u003ca href=\"https://www.mapa.gob.es\"\u003ehttps://www.mapa.gob.es\u003c/a\u003e).\u003c/p\u003e","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-2742457/v1/078d6a63091aed7a374dec9e.png"},{"id":35294580,"identity":"bbee1293-9eac-4c25-8638-b78cbc446201","added_by":"auto","created_at":"2023-04-04 22:08:44","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":36004,"visible":true,"origin":"","legend":"\u003cp\u003eThe bottom panel shows the observed weather data (precipitation, P, in blue, and Evapotranspiration, ET0, in green, in mm) during one of the growth periods studied from December 1, 2002, to July 1\u003csup\u003est\u003c/sup\u003e, 2003. Then (in descending order) the forecast data are shown for a lead time of 7 to 1 months before harvest. For Static Stress calculations with a lead time between 1 and 7 months, the remaining months of the growth period are completed with observed weather data.\u003c/p\u003e","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-2742457/v1/56aea31542c1196f5e989dfc.png"},{"id":35294578,"identity":"320d6350-3732-4526-8f62-a6935eda619d","added_by":"auto","created_at":"2023-04-04 22:08:44","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":17498,"visible":true,"origin":"","legend":"\u003cp\u003ePrediction of winter wheat yield as a function of (A) Static Stress, ζobs, (B) Standardized Precipitation and Evapotranspiration Index (SPEI-3) quarterly aggregated (January-March) and (C) Standardized Precipitation and Evapotranspiration Index (SPEI-3) quarterly aggregated (April-June). Note that the red dot indicates an outlier that was not considered.\u003c/p\u003e","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-2742457/v1/93b123bd8f9d7d9e267d84cb.png"},{"id":35295472,"identity":"2db468da-e06a-42f8-9b8a-f363b5f14e33","added_by":"auto","created_at":"2023-04-04 22:16:44","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":5543,"visible":true,"origin":"","legend":"\u003cp\u003eR\u003csup\u003e2\u003c/sup\u003e values estimated from the regression model to predict crop yield at the end of the growing season for each forecast lead time (From 1 to 7 months lead time; bars in light blue) and the observed data (0 months lead time; bar in dark blue).\u003c/p\u003e","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-2742457/v1/a7b336538dc3eed970a18fe0.png"},{"id":35294581,"identity":"6b6b1e7d-8753-4018-9aa5-a3c07d84b8e9","added_by":"auto","created_at":"2023-04-04 22:08:44","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":24880,"visible":true,"origin":"","legend":"\u003cp\u003eTime series of calculated soil moisture anomalies using weather forecasts for the period 2002-2020.\u003c/p\u003e","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-2742457/v1/b98f43f4897a438c23e0010c.png"},{"id":40512559,"identity":"8360d23a-a2ce-4c2c-816b-10f11945a66e","added_by":"auto","created_at":"2023-07-25 06:22:40","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":674775,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-2742457/v1/47a65085-f687-4a92-8a29-1dd9afbdc9c5.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Early prediction of wheat yield using seasonal weather forecasts and the static stress drought index","fulltext":[{"header":"Introduction","content":"\u003cp\u003eThe ongoing armed conflicts and the climate crisis jeopardize global food security\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. While conflicts are currently viewed as the main driver of food insecurity\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. Extreme climatic events, such as droughts or flooding, are particularly important however as they significantly impact agricultural production\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e,\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e,\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e at a global scale, and increase the risk of armed conflict \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. Weather extremes, and in particular droughts, are expected to become more frequent and severe over the next years due to climate change\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. In Europe, drought-related crop production losses tripled between 1964\u0026ndash;1990 and 1991\u0026ndash;2015: from 2.2\u0026ndash;7.3%\u003csup\u003e9\u003c/sup\u003e. A study by Jenkins\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e showed that in 2003\u0026ndash;2050 annual drought losses will increase by 32% and 57% for short-term and long-term droughts, respectively, in Spain. Spain is a particularly sensitive country to droughts that has been subject to very severe drought events in the last century\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e causing important agricultural losses \u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. Therefore, there is an urgent need to develop tools to anticipate drought impacts and to support the definition of mitigation measures.\u003c/p\u003e \u003cp\u003eDrought is an extreme hydroclimatic event whose definition is not unique\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. The meteorological definition is the most prevalent, \u0026rdquo;a period of more than some particular number of days with precipitation less than some specified small amount\u0026rdquo;\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. However, agricultural drought is not only related to rain scarcity but also poor soil and soil moisture conditions that result in adverse crop responses, such as reduced crop yield\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. Indeed, deep, fertile soils with a large soil water holding capacity constitute a better buffer against droughts than shallow, eroded soils. Further, drought can also deteriorate the soil conditions by increasing soil erodibility and hence soil loss\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. The use of drought indices is the most common approach for quantifying and monitoring drought intensity and impacts on agriculture. Most drought indicators are based on precipitation measurements\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. Among them, the SPI\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e and Standardized Precipitation Evapotranspiration Index (SPEI)\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e are widely used and recommended by the World Meteorological Organization (WMO). However, agricultural droughts, i.e., when crops become affected, usually occur when there is a soil moisture deficit. Jim\u0026eacute;nez-Donaire et al.\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e evaluated the potential of two novel indices, Static Stress, and Dynamic Stress, that consider soil moisture dynamics and found them better correlated with wheat yield than SPI in Southwestern Spain. Other type of drought indicators are the Combined Drought Indicators (CDI) which encompass meteorological, soil moisture and vegetation condition satellite-derived data and are more adequate to monitor agricultural droughts\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. A CDI designed by Sepulcre-Canto et al.\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e distinguishes categorical drought warnings based on satellite-derived data. However, the validation of these categorical classes that represent the severity of the impact of drought, such as crop yield losses, is lacking \u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. The main limitations of CDI are that they need satellite products of high resolution and, especially, that these products are not adequate for future forecasts, and are limited to historical observations to make future predictions.\u003c/p\u003e \u003cp\u003eNumerous studies have linked drought indices to crop yield variability in Spain. Garc\u0026iacute;a-Le\u0026oacute;n et al.\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e concluded that remote sensing data related to vegetation health provided more accurate information on local drought conditions compared to drought indices based solely on rainfall, largely because the former incorporates local biophysical and climate factors inherently. Pe\u0026ntilde;a-Gallardo et al.\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e assessed the impact of drought also on barley and wheat, two rainfed crops in Spain. They found a good correlation between crop yield and SPI and SPEI, however, the weakest correlations were found in South Spain. While these studies show good results in terms of correlating crop yields with drought indices, these results are based on past observed data. To support land management and in particular, the definition of measures to reduce the impact of droughts in agriculture, we should further explore the potential of estimating drought indices from mid-term weather forecasts, i.e. several months in advance. Forecasting droughts that occur during the most sensitive stages of crops is paramount for soil and land management. In this sense, seasonal forecasts can act as a support for stakeholders and policymakers in making more optimal decisions several months in advance. Seasonal forecasts can not only be used to develop drought early warning systems but also for drought impact forecasts \u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e, such as crop yield losses. In general, in the literature there is a lack of studies, evaluating the potential of making future predictions of crop yields using drought indices derived from mid-term (or seasonal) weather forecasts.\u003c/p\u003e \u003cp\u003eIn this study, we present the first study to evaluate the potential of using seasonal weather forecasts combined with a drought index to predict cereal crop yields. Specific objectives are (i) to evaluate the correlation between the Static Stress, based on both precipitation and soil moisture conditions, and compare it against the more traditional indicator SPEI only based on precipitation and (ii) to develop and evaluate a regression model to predict winter wheat yield for a lead time from 0 to 7 months in C\u0026oacute;rdoba (South Spain). The model is based on the use of the Static Stress and seasonal weather forecasts data provided by the ECMWF SEAS5. The information provided here and the proposed methodoly may benefit for farmers and stakeholders to maximize cereal yields and mitigate drought impacts.\u003c/p\u003e"},{"header":"Material And Methods","content":"\u003cp\u003eA schematic overview of the applied methodology is shown in the flowchart in Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e and explained in detail below. The top panel (A) provides a framework for evaluating drought indexes. In this sense, two datasets of observed data were used in this study: observed winter wheat yield (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e box 1), and observed weather data (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e box 2). Observed weather data were used to estimate soil moisture using a soil moisture model (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e box 3). The drought indexes that are estimated using observed weather data are the SPEI, for two different periods, and the Statis Stress Index using soil moisture (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e box 4). Model performance is measured by the coefficient of determination (R\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e) to evaluate which drought index is better correlated to the crop yield (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e box 5). Then, the lower panel (B) provides a framework to define eight linear regression models (one per forecast lead time) using the observed weather data dataset combined with seasonal forecasts (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e box 6) to predict annual wheat yield. Observed values (until the date of issue of the forecast) are merged with seasonal forecasts from 7 months lead time (only seasonal forecast data) to 0 months lead time (only observed data) to cover the winter wheat growing season (December to July) (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e box 7). These 7-month weather data are used to simulate soil moisture (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e. Box 8) and calculate the drought index (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e. Box 9). To evaluate the relationship between observed winter wheat yield and the drought index, calculated as the mean value obtained for the set of 25 members or forecast provided by the ECMWF at each forecast lead time, we computed R\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e box 10). R\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e values are obtained for the 8 lead times that we define previously.\u003c/p\u003e\n\u003cdiv class=\"Section2\" id=\"Sec3\"\u003e\n \u003ch2\u003e1.1. Region of study\u003c/h2\u003e\n \u003cp\u003eThe area of study is the province of C\u0026oacute;rdoba, located in South Spain. It was selected because of its large land extension dominated by rainfed crops, in particular, herbaceous crops such as winter wheat (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eThe climate of this region is Mediterranean, which is characterized by hot and dry summers and cold winters. The average annual temperature is 18\u0026deg;C\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e, with an average annual precipitation of 602 mm, with a standard deviation of 184 mm (1961\u0026ndash;2018). Figure \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eA shows the time series of precipitation during the hydrological year of the study period 2002\u0026ndash;2020 and Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eB the corresponding precipitation for different stages of wheat development: from the end of anthesis (March) to grain filling (April-June). The average precipitation during the selected period was 605 mm, ranging from 314 mm to 1062 mm. The standard deviation was 192 mm.\u003c/p\u003e\n\u003c/div\u003e\n\u003ch2\u003e1.1 . Datasets\u003c/h2\u003e\n\u003cdiv class=\"Section2\" id=\"Sec5\"\u003e\n \u003ch2\u003e1.1.1. Observational crop yield dataset\u003c/h2\u003e\n \u003cp\u003eFor the 2002\u0026ndash;2020 study period winter wheat yield was compiled (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e box 1). The average winter wheat yield was 2657 kg ha\u003csup\u003e\u0026minus;\u0026thinsp;1 28\u003c/sup\u003e in the province of C\u0026oacute;rdoba (dashed line in Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eC). Wheat in this area is usually planted from late November to early December and is usually harvested in late June to early July. In this study, the growing season considered was from December 1st to July 1st .\u003c/p\u003e\n \u003cp\u003eIn Spain, Andalusia accounts for 67.7% of winter wheat production \u003csup\u003e\u003cspan class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e. As shown in Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eC, winter wheat yields have remained relatively stable since the beginning of the study period. However, drops in yields in specific years, such as that of 2005 can be observed. As reported by the Spanish State Meteorological Agency (AEMET), 2005 was the driest year in the study period (1961\u0026ndash;2018)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e, with a mean annual precipitation of 451 mm for peninsular Spain, which represents 70% of the average (640 mm). This severe precipitation deficit together with another drought of unprecedented intensity in the 2007\u0026ndash;2008 hydrological year\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e caused the onset of a long-lasting drought based on SPI-36 from December 2005 to December 2009\u003csup\u003e30\u003c/sup\u003e. It is remarkable that since 2005, the surface area devoted to winter wheat has declined from 140,000 to 40,000 ha (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eD). Espinosa-Tas\u0026oacute;n et al.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e studied the economic impact of multiyear droughts (2005\u0026ndash;2008) in Andalusia. The area dedicated to cereals has decreased since 2005, as well as rainfed and irrigated agricultural production. Nonetheless, some irrigated crops experienced noticeable increases in production in 2005\u0026ndash;2008 due to agricultural expansion since 2000\u003csup\u003e32\u003c/sup\u003e.\u003c/p\u003e\n \u003cdiv class=\"Section3\" id=\"Sec6\"\u003e\n \u003ch2\u003e1.1.2. Observational weather dataset\u003c/h2\u003e\n \u003cp\u003eThe daily weather data (Precipitation, P, and Reference evapotranspiration, ET\u003csub\u003e0\u003c/sub\u003e) during 2002\u0026ndash;2020 was obtained from C\u0026oacute;rdoba station (37.86\u0026deg; N,4.80\u0026deg;W) (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e box 2), at 94 meters above the sea level, of the Agroclimatic Information Network of Andalusia (RIA, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.juntadeandalucia.es/agriculturaypesca/ifapa/riaweb/web/estacion/14/6\u003c/span\u003e\u003c/span\u003e). We used the values proposed by Vanderlinden\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e calculated from the soil map of Andalusia for soil moisture content at wilting point (s\u003csub\u003ew\u003c/sub\u003e) and field capacity (s\u003csub\u003efc\u003c/sub\u003e).\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv class=\"Section3\" id=\"Sec7\"\u003e\n \u003ch2\u003e1.1.3. Forecast dataset\u003c/h2\u003e\n \u003cp\u003eThe seasonal weather forecast data were obtained from the ECMWF SEAS5 (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e box 6). ECMWF provides both real-time seasonal forecasts and historical re-forecasts, i.e. hindcasts, generated with the last version of the system, i.e. SEAS5, currently used for the real-time forecasts. These two products have a 7-month lead time starting the 1st day of the month. They have a spatial resolution of 1⁰ x 1⁰ (36 km approximately) and are composed of 25 members for the hindcast and 50 for the forecast. In this study, daily precipitation and evapotranspiration are retrieved from 2002 until 2016 (hindcasts) and from 2017 to 2020 (forecasts). For the remainder of the study, we will use only the term forecast to refer to both hindcast and forecast.\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv class=\"Section2\" id=\"Sec8\"\u003e\n \u003ch2\u003e1.2. Drought indices\u003c/h2\u003e\n \u003cdiv class=\"Section3\" id=\"Sec9\"\u003e\n \u003ch2\u003e1.2.1. Standardized Precipitation Evaporation Index\u003c/h2\u003e\n \u003cp\u003eThe Standardized Precipitation Evaporation Index (SPEI-n) (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e box 4) is a simple multiscale drought index based on a combination of precipitation and temperature data. SPEI is calculated using a fixed-length moving window based on a monthly difference between precipitation and calculated potential evapotranspiration, using the Hargreaves method based on temperature. In this study, we focus on shorter accumulation periods, or SPEI-3, which is more closely related to the impact of drought on seasonal herbaceous systems\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. According to the Spanish annual crop calendar\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e, the growing season considered for winter wheat was between December and July. SPEI-3 was quarterly aggregated to capture two different growth stages of the plant: mid or anthesis (January-March) and late or grain-filling stages (April-June) as recommended by Garc\u0026iacute;a-Le\u0026oacute;n et al.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. SPEI values range from \u0026minus;\u0026thinsp;3 (extremely dry) to 3 (extremely wet)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. The SPEI values were obtained using the R package SPEI (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://cran.r-project.org/web/\u003c/span\u003e\u003c/span\u003e packages/SPEI/index.html).\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv class=\"Section3\" id=\"Sec10\"\u003e\n \u003ch2\u003e1.2.2. Static Stress\u003c/h2\u003e\n \u003cp\u003eThe Static Stress (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e box 4),\u003cem\u003e\u0026zeta;\u003c/em\u003e, estimates the plant water stress related to soil moisture conditions\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e. It is determined as Eq.\u0026nbsp;(1)\u003csup\u003e17\u003c/sup\u003e :\u003c/p\u003e\n \u003cdiv class=\"Equation\" id=\"Equ1\"\u003e\n \u003cdiv class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e$$\\genfrac{}{}{0pt}{}{\\zeta =0 for s\u0026gt;s*,}{\\zeta =1 for s\u0026lt;{s}_{w}}$$\u003c/div\u003e\n \u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\n \u003c/div\u003e\n \u003cp\u003ewhere the \u003cem\u003e\u0026zeta;\u003c/em\u003e is zero when soil moisture, s, is above the incipient stomatal closure, s*, and is equal to one when soil moisture is below the permanent wilting point, s\u003csub\u003ew\u003c/sub\u003e. For soil moisture values between s* and s\u003csub\u003ew\u003c/sub\u003e, the general expression is as Eq.\u0026nbsp;(\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e):\u003c/p\u003e\n \u003cdiv class=\"Equation\" id=\"Equ2\"\u003e\n \u003cdiv class=\"mathdisplay\" id=\"FileID_Equ2\" name=\"EquationSource\"\u003e$$\\zeta \\left(t\\right)={\\left[\\frac{{s}^{*}-s\\left(t\\right)}{{s}^{*}-{s}_{w}}\\right]}^{q}, for {s}_{w}\\le s\\left(t\\right)\\le {s}^{*},$$\u003c/div\u003e\n \u003cdiv class=\"EquationNumber\"\u003e2\u003c/div\u003e\n \u003c/div\u003e\n \u003cp\u003ewhere q is a measure of the non-linearity of the effects of soil moisture deficit on plant conditions. In this study, we use a value of 1 which implies a linear relation between static water stress and soil moisture deficit\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\n \u003cp\u003eSoil moisture dynamics are calculated following the same approach presented by Jim\u0026eacute;nez-Donaire et al.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e which consists of a soil water-balance \u0026ldquo;bucket\u0026rdquo; model considering infiltration, evapotranspiration, and deep seepage over the root zone (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e box 3 and box 8).\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv class=\"Section2\" id=\"Sec11\"\u003e\n \u003ch2\u003e1.3. Regression model for wheat yield\u003c/h2\u003e\n \u003cp\u003eFor each forecast lead time, we used a linear regression model to estimate the crop yield at the end of the growing season. For this purpose, we define 8 scenarios using both observed and forecast data ( Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e box 6 and 7 and Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e top to bottom panels). We evaluated the performance of the model to predict the crop yield by the end of the harvest period (1 July) for 7 forecast lead times: from 7 months (starting 1 December) to 1 month (starting 1 June). When the starting month is different from December, e.g., February, observed precipitation and evapotranspiration observed values are used until the time of forecast initialization, e.g. 1 February. Then, for each lead time, we calculated the mean of the selected drought index for the 25 forecast ensemble members (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e). The process is repeated each month until estimations are obtained for a lead time of 0 months, i.e., only observed data (bottom panel in Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e). Since we are interested in predicting drought events, no bias correction has been applied to the seasonal weather forecast. The justification for this is that bias correction pushes forecasts to be more like average past weather, therefore bias correction may reduce the \u0026ldquo;good signal\u0026rdquo; that may be present in the original forecast in months that will indeed be significantly drier than the average\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\n \u003cp\u003eThe correlation is calculated between the forecasted yearly mean value of the drought index and the observed winter wheat yields for lead times between 0 and 7 months. For each lead time, the R\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e value will be presented in the form of a bar plot.\u003c/p\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n\u003c/div\u003e"},{"header":"Results And Discussion","content":"\u003cdiv class=\"Section2\" id=\"Sec13\"\u003e\n \u003ch2\u003e2.1. Evaluation of Static Stress for Prediction of Wheat Yield\u003c/h2\u003e\n \u003cp\u003eObserved Static Stress and SPEI-3 were evaluated and compared based on their capacity to explain winter wheat yield variability (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e box 5). The results of the three statistical models with the winter wheat yield data and these drought indices are presented in Fig. \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003e. Static Stress shows a good correlation with yield, with an R\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e of 0.57. By comparison, SPEI-3 performs poorly, with an R\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e value of 0.25 and 0.12 for the mid and late stages respectively. It should be noted that 2018 (Fig. \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003e. in red) was identified as an outlier and removed from the analysis. Based on the findings of this study, losses in wheat yield can be well explained by the Static Stress for the study period. While this agrees with the results of other studies\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e other factors may also influence yield and hence, the occurrence of a drought event may not necessarily lead to a reduction in yield \u003csup\u003e\u003cspan class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e. This might explain why in 2018 the high crop yield does not correspond to a low Static stress value. Nevertheless, exceptionally intense droughts can considerably increase the probability of yield loss for wheat\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e. Another important difference between Static Stress and SPEI is that the former only indicates the presence of drought stress. Therefore, wet years are characterized by 0 values. In contrast, the relationship between yield and SPEI is overall positive, but especially for Jan-Mar, the relation is more complex, and first positive up till 0 and then negative. This can be explained because dryer years are characterized by more negative values, but wetter than average years can also see reductions in crop yield which explains the decline of yields for values\u0026thinsp;\u0026gt;\u0026thinsp;0. The highest yields are generally observed for years with precipitation around average, i.e. SPEI-3(Jan-Mar)\u0026thinsp;=\u0026thinsp;0. For SEPI-3(Apr-Jun), similar behavior can be observed but the trend is less clear. Overall, the results of this study case show that Static Stress is a better predictor of crop yield than SPEI and indicate that drought indicators that consider both precipitation and soil moisture conditions are better predictors of crop yields than indicators that only consider precipitation.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv class=\"Section2\" id=\"Sec14\"\u003e\n \u003ch2\u003e2.2. Forecast of wheat yield\u003c/h2\u003e\n \u003cp\u003eModel performance measured by R\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e value per lead-time, based on the mean Static stress calculated using the set of 25 members of the seasonal weather forecasts, and observed past weather data, to predict winter wheat yield is assessed (Fig. \u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003e) for each forecast lead time. As expected, R\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e values gradually decrease with longer forecast lead times. In general, the proposed linear regression models show good performance (R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.5; p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05) for lead times equal to or less than 4 months, i.e. from March to July. Figure \u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003e shows that the best correlation is obtained for a forecast lead time of 2 months (R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.64; p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05), i.e. combination of 5 months of observed weather data (December-April) and 2 months of seasonal weather forecast data (May-June). The results for a 3 months lead time (R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.58) are similar to those obtained for a 0-month lead time, i.e. using only observed data (R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.57). There is a significant difference between the results obtained for 5 months (R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.377), i.e. from February to July, and 6 months (R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.17), i.e. from January to July. Surprisingly, the correlation is higher for a 2-month lead time than for a 1-month lead time and even when using only observed data (distinguished in Fig. \u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003e in dark blue; 0-month lead time).\u003c/p\u003e\n \u003cp\u003eThis study proposes a simple method to predict yields through a regression method with relatively low data and computational requirements. In the literature, there is a wide range of models used for yield forecasting but regression models showed better results in 74% of studies assessed by Schauberger et al.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e. Meroni et al.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e also claimed that regression models outperformed 60% of tested machine learning models in predicting barley, soft, and durum wheat yields in Algeria. Moreover, machine-learning models require larger datasets\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\n \u003cp\u003eTo our knowledge, we present the first study that assesses the feasibility of combining drought indices with seasonal forecasts to predict crop yields. A study by Bento et al.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e used linear regression models and seasonal precipitation and temperature forecasts, instead of drought indices, from SEAS5 to predict wheat and barley yield in the Iberian Peninsula. Similarly to this study, winter wheat yield forecasts only started to gain skill from April onwards, i.e. three months before the harvest. They remark that some variables chosen by their selected regression models are bound to change in the coming decades due to unavoidable warning scenarios and need to be parameterized locally. In Europe, Ceglar and Toreti\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e tested the predictability of drought events during winter wheat production however they did not relate drought indices with crop yield. In their study, skillful and reliable predictions of drought events using SEAS5 forecast with indicators based on SPEI can be made already at the end of winter. In Australia, Jin et al.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e found significant yield forecast skill improvement at the beginning of the winter wheat cropping season, such as 1st May or 1st June by using the seasonal climate forecast model ACCESS-S1 and a process-based crop model. Operational yield forecasting systems exist at a regional scale, such as the European Union (Mars\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e), assessing the increasingly variable climatic conditions that are impacting crop production systems throughout the EU, and on a global scale\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e. Despite the challenges associated with implementing a global forecasting system, accurate information could be beneficial to agribusinesses and food security in regions where crop yields are susceptible to climate change and food markets are not connected to international trade\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e. Next-generation crop yield forecasting systems are increasingly sophisticated and include machine learning techniques featuring neural networks\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e. Khaki and Wang\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e used deep neural networks for forecasting crop yield pointing out the sensitivity of the model to weather data origin. They also highlighted that even when their model outperformed other models, it is more accurate than explanatory.\u003c/p\u003e\n \u003cp\u003eIn Spain, efforts need to be made to make such information available to decision-makers in a form that is understandable and useable. The use of in-season decision support systems that incorporate crop models might improve the link between agricultural practices and climate information\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e. In several countries, to reduce the risk of crop failure and increase farmers\u0026apos; profits, seasonal climate forecasts are linked to crop simulation models to provide potentially useful information for stakeholders. Since 2019, the commercial platform CropProphet used ECMWF forecast data for their crop yield and production forecast models, in the USA. Their resulting 15-day yield forecast models provide the advantage of being ahead of market expectations and futures market trading\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e. Furthermore, another platform, Planet, supplies daily corn yield forecasts of the growing season also in the USA. This is done by using an algorithm that makes optimal use of archival records, giving their statistical models an unparalleled trove of data to correlate with historical yields\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e. In Australia, Wheatcast\u0026trade; integrated analytics systems lead to a yield forecast at any time and scale\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\n \u003cp\u003eWhile this study show the potential for the practical application of the proposed method some limitations of the results must be considered. Winter wheat yields correspond to the average of the whole province of Cordoba, not to specific areas. With higher spatial resolution yield data we could have better link changes in yield due to local changes in weather and soil moisture. Because of this, we also used the same soil and plant parameters in the soil water balance model and to calculate plant stress were unique.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"Conclusions","content":"\u003cp\u003eIn this study, we assessed the potential of using seasonal weather forecasts and the Static Stress drought index to predict winter wheat yield in South Spain. First, we find that Static Stress is better correlated to crop yield than SPEI and that Static Stress derived from seasonal forecasts shows a good performance in predicting crop yields for 4 or fewer months before harvest, i.e., from March to July. These results indicate that drought indicators that consider soil moisture conditions are better predictors of crop yields than indicators that only consider precipitation. Furthermore, this study demonstrates the potential of using the Static Stress index derived from seasonal forecasts to predict crop yields. Knowing the yield before harvest can help stakeholders maximize production while avoiding expenses that might affect small farmers and economies.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cdiv class=\"Section2\" id=\"Sec13\"\u003e\n \u003cp\u003e\u003cstrong\u003eData availability Statement\u003c/strong\u003e. The datasets generated during and/or analyzed during the current study are available from the corresponding author on request.\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the Spanish Ministry of Science, Innovation, and Universities co-financed by the EU through the project \u0026ldquo;Quantifying the impact of soil erosion and climate change on soil security by using alternative fallout radionuclides\u0026rdquo; (PID2019-109924RB-I00/AEI/10.13039/501100011033). Garc\u0026iacute;a-Gamero, Pe\u0026ntilde;uela, and Vanwalleghem acknowledge financial support from the Department of Agronomy, of the Spanish Ministry of Science and Innovation, the Spanish State Research Agency, through the Severo Ochoa and Mar\u0026iacute;a de Maeztu Program for Centers and Units of Excellence in R\u0026amp;D (Ref. CEX2019-000968-M).\u0026nbsp;The authors wish to thank the Copernicus Climate Change and Atmosphere Monitoring Services for providing the seasonal forecasts generated by the ECMWF seasonal forecasting systems (SEAS5). Neither the European Commission nor ECMWF is responsible for any use that may be made of the Copernicus information or data it contains.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eVGG and TV designed the research. \u0026nbsp;VGG and \u003csup\u003e1\u003c/sup\u003eAP review the literature and collected the data. VGG, \u003csup\u003e1\u003c/sup\u003eAP, and TV designed the analytical strategy. \u0026nbsp;VGG conducted the data analyses and produced the results. VGG performed the interpretation of the results with the contribution of \u003csup\u003e1\u003c/sup\u003eAP, \u003csup\u003e2\u003c/sup\u003eAP and TV. VGG wrote a first draft of the manuscript and thereafter \u003csup\u003e1\u003c/sup\u003eAP, \u003csup\u003e2\u003c/sup\u003eAP\u003csup\u003e,\u003c/sup\u003e and TV made subsequent revisions to the original text and figures.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAdditional information\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eScheffran, J. \u0026amp; Battaglini, A. Climate and conflicts: the security risks of global warming. Reg Env. Chang. (2011). doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10113-010-0175-8\u003c/span\u003e\u003cspan address=\"10.1007/s10113-010-0175-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFood Security Information Network. \u003cem\u003eGLOBAL REPORT ON FOOD CRISES\u003c/em\u003e. (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhu, P. \u003cem\u003eet al.\u003c/em\u003e Warming reduces global agricultural production by decreasing cropping frequency and yields. Nat. Clim. Chang. 12, (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDasgupta, S. \u0026amp; Robinson, E. J. Z. Attributing changes in food insecurity to a changing climate. Sci. Rep. 1\u0026ndash;11 (2022). doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41598-022-08696-x\u003c/span\u003e\u003cspan address=\"10.1038/s41598-022-08696-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKrishnamurthy R, P. K., Fisher, J. B., Choularton, R. J. \u0026amp; Kareiva, P. M. Anticipating drought-related food security changes. Nat. Sustain. (2022). doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41893-022-00962-0\u003c/span\u003e\u003cspan address=\"10.1038/s41893-022-00962-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGe, Q. \u003cem\u003eet al.\u003c/em\u003e Modelling armed conflict risk under climate change with machine learning and time-series data. Nat. Commun. 13, 1\u0026ndash;8 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVon Uexkull, N., Croicu, M., Fjelde, H. \u0026amp; Buhaug, H. Civil conflict sensitivity to growing-season drought. \u003cem\u003eProc. Natl. Acad. Sci. U. S. A.\u003c/em\u003e 113, 12391\u0026ndash;12396 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHasegawa, T. \u003cem\u003eet al.\u003c/em\u003e Extreme climate events increase risk of global food insecurity and adaptation needs. Nat. Food 2, 587\u0026ndash;595 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArmada Br\u0026aacute;s, T., Seixas, J., Carvalhais, N. \u0026amp; J\u0026auml;germeyr, J. Severity of drought and heatwave crop losses tripled over the last five decades in Europe. Environ. Res. Lett. 16, 65012 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJenkins, K. Indirect economic losses of drought under future projections of climate change: A case study for Spain. Nat. Hazards 69, 1967\u0026ndash;1986 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGonz\u0026aacute;lez-Hidalgo, J. C. \u003cem\u003eet al.\u003c/em\u003e High-resolution spatio-temporal analyses of drought episodes in the western Mediterranean basin (Spanish mainland, Iberian Peninsula). Acta Geophys. 66, 381\u0026ndash;392 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVicente-Serrano, S. M. Spatial and temporal analysis of droughts in the Iberian Peninsula (1910\u0026ndash;2000). Hydrol. Sci. J. 51, 83\u0026ndash;97 (2006).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUPA. \u003cem\u003eAgricultura familiar en Espa\u0026ntilde;a. Anuario 2017\u003c/em\u003e. (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWilhite, D. A. \u0026amp; Glantz, M. H. Understanding the Drought Phenomenon: The Role of Definitions. Water Int. 10, 111\u0026ndash;120 (1985).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGreat Britain Meteorological Office. \u003cem\u003eMeteorological Glossary\u003c/em\u003e. \u003cem\u003eGreat Britain Meteorological Office\u003c/em\u003e (Chemical Publishing Co., 1951). doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/j.1477-8696.1964.tb02532.x\u003c/span\u003e\u003cspan address=\"10.1002/j.1477-8696.1964.tb02532.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKang, Y., Khan, S. \u0026amp; Ma, X. Climate change impacts on crop yield, crop water productivity and food security - A review. Prog. Nat. Sci. 19, 1665\u0026ndash;1674 (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePorporato, A., Laio, F., Ridolfi, L. \u0026amp; Rodriguez-Iturbe, I. Plants in water-controlled ecosystems: active role in hydrologic processes and response to water stress III. Vegetation water stress. Adv. Water Resour. 24, 725\u0026ndash;744 (2001).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaumhardt, R. L. \u0026amp; Blanco-Canqui, H. Soil: Conservation Practices. Encycl. Agric. Food Syst. 5, 153\u0026ndash;165 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDutra, E. \u003cem\u003eet al.\u003c/em\u003e Global meteorological drought-Part 1: Probabilistic monitoring. Hydrol. Earth Syst. Sci. 18, 2657\u0026ndash;2667 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBhanu Kumar, O. S. R. U., Ramalingeswara Rao, S., Ramaganathan, S. \u0026amp; Raju, S. S. Role of Intra-seasonal Oscillations on Monsoon Floods and Droughts over India. Asia-Pacific J. Atmos. Sci 46, 21\u0026ndash;28 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVicente-Serrano, S. M., Beguer\u0026iacute;a, S. \u0026amp; L\u0026oacute;pez-Moreno, J. I. A multiscalar drought index sensitive to global warming: The standardized precipitation evapotranspiration index. J. Clim. 23, 1696\u0026ndash;1718 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJim\u0026eacute;nez-Donaire, M. del P., Gir\u0026aacute;ldez, J. V. \u0026amp; Vanwalleghem, T. Evaluation of Drought Stress in Cereal through Probabilistic Modelling of Soil Moisture Dynamics. Water 12, 2592 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSepulcre-Canto, G., Horion, S., Singleton, A., Carrao, H. \u0026amp; Vogt, J. Development of a Combined Drought Indicator to detect agricultural drought in Europe. Nat. Hazards Earth Syst. Sci. 12, 3519\u0026ndash;3531 (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZargar, A., Sadiq, R., Naser, B. \u0026amp; Khan, F. I. A review of drought indices. Environ. Rev. 19, 333\u0026ndash;349 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGarc\u0026iacute;a-Le\u0026oacute;n, D., Contreras, S. \u0026amp; Hunink, J. Comparison of meteorological and satellite-based drought indices as yield predictors of Spanish cereals. Agric. Water Manag. 213, 388\u0026ndash;396 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePe\u0026ntilde;a-gallardo, M., Vicente-serrano, S. M., Dom\u0026iacute;nguez-castro, F. \u0026amp; Beguer\u0026iacute;a, S. The impact of drought on the productivity of two rainfed crops in Spain. Nat. Hazards Earth Syst. Sci. 1215\u0026ndash;1234 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSutanto, S. J., van der Weert, M., Wanders, N., Blauhut, V. \u0026amp; Van Lanen, H. A. J. Moving from drought hazard to impact forecasts. Nat. Commun. 10, 1\u0026ndash;7 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMinisterio de Agricultura Pesca y Alimentaci\u0026oacute;n. Encuesta sobre Superficies y Rendimientos Cultivos (ESYRCE). \u003cem\u003eBolet\u0026iacute;n Mensual de Estad\u0026iacute;stica\u003c/em\u003e (2022). Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.mapa.gob.es/es/estadistica/temas/estadisticas-agrarias/agricultura/esyrce/\u003c/span\u003e\u003cspan address=\"https://www.mapa.gob.es/es/estadistica/temas/estadisticas-agrarias/agricultura/esyrce/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. (Accessed: 15th October 2022)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMinisterio de Agricultura Pesca y Alimentaci\u0026oacute;n. \u003cem\u003eEstad\u0026iacute;stica Anual De Superficies y Producciones de cultivos\u003c/em\u003e. (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRomero Fresneda, R. \u003cem\u003eet al. Comportamiento de las precipitaciones en Espa\u0026ntilde;a y periodos de sequ\u0026iacute;a (periodo\u003c/em\u003e 1961\u003cem\u003e\u0026ndash;2018)\u003c/em\u003e. \u003cem\u003eComportamiento de las precipitaciones en Espa\u0026ntilde;a y periodos de sequ\u0026iacute;a (periodo 1961\u0026ndash;2018)\u003c/em\u003e (Agencia Estatal de Meteorolog\u0026iacute;a, 2020). doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.31978/666-20-006-0\u003c/span\u003e\u003cspan address=\"10.31978/666-20-006-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMestre, A. \u0026amp; Moreno, J. V. El Seguimiento de la Sequ\u0026iacute;a en Eastaciones y Sobre Grandes Cuencas a Trav\u0026eacute;s de Uso del \u0026iacute;ndice SPI. Bolet\u0026iacute;n AME 36\u0026ndash;41 (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEspinosa-Tas\u0026oacute;n, J., Berbel, J., Guti\u0026eacute;rrez-Mart\u0026iacute;n, C. \u0026amp; Musolino, D. A. Socioeconomic impact of 2005\u0026ndash;2008 drought in Andalusian agriculture. Sci. Total Environ. 826, (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVanderlinden, K. AN\u0026Aacute;LISIS DE PROCESOS HIDROL\u0026Oacute;GICOS A DIFERENTES ESCALAS ESPACIO-TEMPORALES. (UNIVERSIDAD DE C\u0026Oacute;RDOBA, 2001).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOteros, J., Garc\u0026iacute;a-Mozo, H., Botey, R., Mestre, A. \u0026amp; Gal\u0026aacute;n, C. Variations in cereal crop phenology in Spain over the last twenty-six years (1986\u0026ndash;2012). Clim. Change 130, 545\u0026ndash;558 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePe\u0026ntilde;uela, A., Hutton, C. \u0026amp; Pianosi, F. Assessing the value of seasonal hydrological forecasts for improving water resource management: Insights from a pilot application in the UK. Hydrol. Earth Syst. Sci. 24, 6059\u0026ndash;6073 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLeng, G. \u0026amp; Hall, J. Crop yield sensitivity of global major agricultural countries to droughts and the projected changes in the future. Sci. Total Environ. 654, 811\u0026ndash;821 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchauberger, B., J\u0026auml;germeyr, J. \u0026amp; Gornott, C. A systematic review of local to regional yield forecasting approaches and frequently used data resources. Eur. J. Agron. 120, 126153 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMeroni, M., Waldner, F., Seguini, L., Kerdiles, H. \u0026amp; Rembold, F. Yield forecasting with machine learning and small data: What gains for grains? Agric. For. Meteorol. 308\u0026ndash;309, (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBento, V. A. \u003cem\u003eet al.\u003c/em\u003e Persistence versus dynamical seasonal forecasts of cereal crop yields. Sci. Reports 2022 \u003cem\u003e121\u003c/em\u003e 12, 1\u0026ndash;11 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCeglar, A. \u0026amp; Toreti, A. Seasonal climate forecast can inform the European agricultural sector well in advance of harvesting. Clim. Atmos. Sci. (2021). doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41612-021-00198-3\u003c/span\u003e\u003cspan address=\"10.1038/s41612-021-00198-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJin, H., Li, M., Hopwood, G., Hochman, Z. \u0026amp; Bakar, K. S. Improving early-season wheat yield forecasts driven by probabilistic seasonal climate forecasts. Agric. For. Meteorol. 315, 108832 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan der Velde, M. \u0026amp; Nisini, L. Performance of the MARS-crop yield forecasting system for the European Union: Assessing accuracy, in-season, and year-to-year improvements from 1993 to 2015. Agric. Syst. 168, 203\u0026ndash;212 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDoi, T., Sakurai, G. \u0026amp; Iizumi, T. Seasonal Predictability of Four Major Crop Yields Worldwide by a Hybrid System of Dynamical Climate Prediction and Eco-Physiological Crop-Growth Simulation. Front. Sustain. Food Syst. 4, (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKhaki, S. \u0026amp; Wang, L. Crop Yield Prediction Using Deep Neural Networks. Front. Plant Sci. 1, 621 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuarin, J. R. \u0026amp; Asseng, S. \u003cem\u003eWheat crop modelling to improve yields\u003c/em\u003e. (Burleigh Dodds Science Publishing Limited, 2017). doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.19103/as.2016.0004.27\u003c/span\u003e\u003cspan address=\"10.19103/as.2016.0004.27\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDutton, J. Corn Yield Forecast: The #1 Critical Equation | CropProphet. (2020). Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.cropprophet.com/corn-yield-forecast-model/\u003c/span\u003e\u003cspan address=\"https://www.cropprophet.com/corn-yield-forecast-model/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. (Accessed: 26th October 2022)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan der Schalie, R. CornCaster: How Planet\u0026rsquo;s Yield Forecasting Solution Is Helping Agriculturists And Economists Get Ahead Of This Year\u0026rsquo;s Harvest. (2022). Available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.planet.com/pulse/corncaster-how-planets-yield-forecasting-solution-is-helping-agriculturists-and-economists-get-ahead-of-this-years-harvest/\u003c/span\u003e\u003cspan address=\"https://www.planet.com/pulse/corncaster-how-planets-yield-forecasting-solution-is-helping-agriculturists-and-economists-get-ahead-of-this-years-harvest/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. (Accessed: 29th October 2022)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLawes, R. \u003cem\u003eet al.\u003c/em\u003e Graincast\u0026trade;: monitoring crop production across the Australian grainbelt. Crop Pasture Sci. (2022). doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1071/cp21386\u003c/span\u003e\u003cspan address=\"10.1071/cp21386\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-2742457/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-2742457/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eCrop yield prediction considering soil moisture as a proxy for water supply remains crucial for global food security. This study evaluates the potential of using seasonal weather forecasts combined with a drought index, Static Stress, based on both precipitation and soil moisture conditions to predict winter wheat yield 7 to 1 month in advance in C\u0026oacute;rdoba (South Spain). First, using observed climate and crop yield data we evaluate the use of Static Stress, as a potential crop yield predictor and compare it to a more traditionally used index, the SPEI, which is only based on precipitation conditions. Then we evaluate the performance of simple linear regression models to predict crop yields from forecasted Static Stress values calculated using weather forecast data from the ECMWF seasonal forecasting system (SEAS5). We find that Static Stress is better correlated to crop yield than SPEI and that Static Stress derived from seasonal forecasts has a good performance (R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.5; p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05) for crop yield predictions of 4 or fewer months before harvest, i.e., from March to July. In this case study, these results indicate that drought indicators that consider soil moisture conditions are better predictors of crop yields than indicators that only consider precipitation. Furthermore, this study demonstrates the potential of using simple regression models together with mid-term forecasts of the Static Stress index to maximize cereal yields and mitigate drought impacts.\u003c/p\u003e","manuscriptTitle":"Early prediction of wheat yield using seasonal weather forecasts and the static stress drought index","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2023-04-04 22:08:39","doi":"10.21203/rs.3.rs-2742457/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"34f1a023-09f5-4c12-a357-521e0839c2cf","owner":[],"postedDate":"April 4th, 2023","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":20378468,"name":"Earth and environmental sciences/Hydrology"},{"id":20378469,"name":"Earth and environmental sciences/Climate sciences/Climate change"},{"id":20378470,"name":"Earth and environmental sciences/Climate sciences/Hydrology"}],"tags":[],"updatedAt":"2023-07-25T06:14:32+00:00","versionOfRecord":[],"versionCreatedAt":"2023-04-04 22:08:39","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-2742457","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-2742457","identity":"rs-2742457","version":["v1"]},"buildId":"rHA-KDH7Qsr4HCuvH75dn","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}