Climate Change Effects on Texas Dryland Winter Wheat Yields

Abstract

Wheat offers winter forage for cattle grazing and is one of the most valuable cash crops in Texas. In this study, we evaluate the impacts of climate change projections on winter wheat grain yields in five major wheat producing counties in Texas (Deaf Smith, Ochiltree, Hansford, Moore, and Parmer). For this purpose, extant soil and climate data were utilized in conjunction with Agricultural Policy Environmental eXtender (APEX) and Coupled Model Intercomparison Project—Phase 5 (CMIP 5) climate projections to determine the most reasonable future trajectory of Texas winter wheat yields. The results indicate that Deaf Smith and Parmer counties are projected to experience the greatest yield decrease, 33.33%, about 696 kg/ha under the CMIP5 RCP4.5 (Texas projected temperature increase between 2.2 and 3.3 °C) 2046–2070 scenario compared to a 1981–2017 baseline. The maximum percentage yield increase was noticed in Ochiltree County under the CMIP5 RCP8.5 2071–2095 scenario, with an 84.2% (about 1857 kg/ha) yield increase compared to the 1981–2017 baseline. Parmer County is projected to experience the greatest yield decrease of 20%, about 348 kg/ha, under the RCP4.5 2046–2070 scenario when compared to the 1981–2005 baseline. The maximum percentage yield increase is projected for Ochiltree County—a 105.9% increase, about 2089 kg/ha—under the RCP8.5 2071–2095 scenario when compared to the 1981–2005 baseline. In general, with few exceptions, winter wheat yields are projected to rise under the projected climate scenarios.

1. Introduction

In the United States, winter wheat is the third largest crop, accounting for 70% to 80% of production in terms of acreage and value. A 51% reduction in high-yielding and more favorable wheat production areas has been observed worldwide as a result of heat stress from climate change [1]. During the 2022–2023 marketing year, 43.5 billion kg of durum, winter, and other spring wheat were produced by farmers in the U.S. from a harvested area of 14.4 million ha [2].

Texas generally has wheat acreage of the hard-red winter class and, due to new variety developments, which include improvements to disease resistance, an increased expansion in acreage in South and Central Texas has been noticed [3]. The High Plains region in Texas (THP) accounted for 47% of the wheat production in the United States between 1965 and 2010. From 2007 to 2017, 1.1 million ha of winter wheat were grown annually in this region [4]. Texas wheat farmers have planted between 2.6 million and 2.2 million ha from 2000 to 2015. These values have decreased to 1.8 million and 2.2 million ha in the last three years. Wheat production totaled 914.5 million kg in 2006, the lowest point between 2000 and 2020 [5]. In the following year, 2007, production had a significant increase to 3.8 billion kg [5]. This value was the highest production total during the twenty-year span. Throughout this high-yielding production year, wheat averaged about 2488.3 kg per ha [5]. Wheat production was 1.9 and 1.7 billion kg in 2019 and 2020, respectively [5,6].

In 2006, weather conditions were considered poor to extremely poor, causing a negative impact on crop production that year [7]. Texas was deemed to be in a tough drought that was having a major effect on the wheat crops. Extreme drought combined with elevated temperatures caused a significant amount of damage to the wheat fields because these weather conditions are not suitable for wheat growth [8]. The 2011 drought in Texas had a heavy impact on winter wheat. Prior to the drought, the acreage of winter wheat increased in the southern climate zones and decreased in the northern climate zones [9]. These extreme and poor climate conditions are a major contributing factor to why the total production for 2006 was the lowest point within the twenty-year span of 2000 to 2020 [5].

Under the management practices of the current land use and the worsening climate conditions, the THP (Texas High Plains) will not be able to mitigate the drying conditions and maintain its agricultural productivity [10]. More recently, the increase in acres planted is due to technological advances in equipment and studies that have developed improved varieties of wheat [11].

The impact of climate change on global food prices, such as grain, has been examined to determine the relationship. Some indicate a minimal change or a slight decrease in the prices of grain up to an increase in global temperatures of 3 °C, after which prices will begin to increase as production decreases [12]. Climate change variability occurred in 2006 in multiple parts of the world when extreme weather led to decreases in global cereal production. The decrease in production yields was at fault for the increase in world food prices [12]. Studies show the demand for agricultural products will increase by 70–110 percent by 2050 [13].

Changes in climate such as precipitation, temperature, and increases in atmospheric CO2 affect crop production [14]. Increases in temperature have a more negative impact on crop yields [15]. Climate change is inevitable and has an impact on various industries, including agriculture. As agriculture is a vital source for our global food security, it is important to identify the effects of climate change on this industry [16]. Cui (2020) [17] evaluated the temporal variation of averages for weather over decades to estimate crop acreage elasticity with respect to climate change within the United States, providing an empirical estimate of acreage elasticity with regard to climate by analyzing the effects of planted acres for soybeans and corn on climate change in the long run. Over the last century, records show that the mean annual temperatures for areas where maize, rice, wheat, and soybeans are grown have increased by 1 °C. These temperatures are expected to continuously increase in the future [18]. In the United States, wheat yield loss was −5.5 ± 4.4%/°C [18]. Asseng et al. (2013) [19] describe a correlation between crop yields and a strong dependency on temperature, especially when temperature increases exceed 3 °C combined with atmospheric CO₂ variations. The elevated temperatures and CO₂ had impacts on crop growth. Trends in rising temperatures have the potential to impact winter wheat crop growth, development, and yield in the United States’ Southern High Plains region of Texas. As a result, adopting risk-mitigation measures is critical for future success [1].

Evaluations of wheat production despite geoclimatic conditions from 1839 to 2007 showed that farmers could adjust to continue producing wheat. Some of those adjustments included location changes and technological advancements. Wheat output in the United States increased 26-fold between 1839 and 2009 [20]. Innovative technology and adaptation strategies have allowed farmers to produce wheat in new environments that were once considered too harsh in North America and should continue to be implemented to combat the negative effects of wheat production [21]. Biological innovations are also vital in assisting expansion where temperatures increase and precipitation decreases [20].

The Agricultural Policy Environmental eXtender (APEX) model [Williams et al., 2000], which is utilized in this study, was created based on the EPIC model. This multifield model was created to address environmental problems related to livestock and agricultural production systems on a whole farm or small watershed basis [22]. Over time, APEX has continued to evolve and be used for a broad range of environmental assessments [23]. The calibrated APEX simulation model is able to determine the impacts of various agricultural management strategies on crop yields. Operating on a daily time step, this physically based model has the capability of generating long-term simulations at the small watershed or whole farm scale [22,24,25,26,27]. Input data used within the APEX model includes soil, daily weather, management strategies, crop type, watershed characteristics, and site location [28]. APEX is calibrated against county-level crop yield annual data, which can be obtained through USDA-NASS, and historical weather and agricultural production data. The key outputs from the model are crop yields, sediment and edge-of-field nutrient losses, and other water and nutrient balance indicators [29,30].

The literature relevant to this study focuses on the following topics: climate change impact on agriculture; suitability of agricultural land under climate change; impacts of climate change on crop production; mitigation strategies.

Olmstead et al. (2010) [20] and Zabel et al. (2014) [13] describe how climate change in Northern America is projected to impact major grain-growing areas in this region, causing issues by altering the suitability of growing areas for wheat. This may result in crop yields and production not being as high as expected due to the changes in climate, ultimately having a negative impact on farmers’ income [31]. The annual mean temperature in Northern America is projected to increase by about 2 to 3 °C in coastal regions, over 5 °C in more northern latitudes, and 3 to 4 °C in major grain-growing regions by 2100. Wheat should have displayed a yield decrease in a study by Sloat et al. (2020) [21] but, as a result of migrating and an increase in irrigation, rainfed wheat is now growing in areas with temperatures that are more favorable.

Future crop yields will be greatly dependent on extreme weather events and changes in climate have already begun to show impacts on agricultural productivity, which could potentially hinder the availability and quality of food [16,32,33,34]. Osei et al. (2023) and Talebizadeh et al. (2018) [29,35] utilized the APEX model to evaluate the effects of management strategies on ecological and agronomic indicators like crop yields and growth, water quality, and livestock grazing. The key Apex model outputs included crop yields, sediment and edge-of-field nutrient losses, and other nutrient and water balance indicators.

The current study contributes to the literature in several ways. First, an analysis of Texas wheat yields is over time to review changes in yields during the study time period, which allows this study to track trends in yields compared to climate, crop migration, and mitigation strategies used. Data is broken into five counties to evaluate county-level yields and averages, along with coordinates from highly active growing areas from the PRISM to gather weather data. Second, the study uses the APEX and CMIP5 models to simulate and project wheat yields in the five counties, which was verified against the county level data from USDA NASS. The projections offer insight into each climate scenario and its impacts on wheat yields. This study aimed to understand the effects of climate variability on Texas wheat yields and aids in identifying the possible changes to help develop mitigation strategies to protect global food security. The objective of this research is to analyze the impacts of climate change on winter wheat production in Texas. Specifically, the study will include the following objectives: to analyze historical wheat yield, production, and price; to evaluate the risk analysis of wheat yield under different APEX scenarios; and to evaluate potential yields under CMIP5 climate change scenarios.

The remainder of this study is organized into the following sections: Section 2 explains the data and methods used in the study. The results of the study are presented in Section 3. Finally, Section 4 contains the conclusion and discussion of this research.

2. Materials and Methods

2.1. Data

Production data for this study were collected from the USDA National Agriculture Statistics Service (NASS) from 1968 to 2020. Wheat production, acres planted, yield, and price for the marketing year were collected for the annual average for Texas. In addition, county-level data on wheat production, acres planted, and yield for Texas were collected as well. Wheat prices, yield, production, acres planted, acres harvested, and fertilizer rates and prices were gathered from the USDA NASS and ERS (Economic Research Service) in Texas. Some of the collected data were used to validate the results from the APEX model.

Fertilizer data for wheat in Texas includes nitrogen, phosphate, and potash. The data include the percentage of Texas wheat acreage receiving each fertilizer and the rate per fertilized acre receiving each of the three fertilizers. Fertilizer data were gathered from Texas AgriLife from 1964 to 2017, which was last updated in 2019. These data were used in the management file for the APEX model and were used to analyze the trend between the amount of fertilizer used and the yields.

Weather data collected for the study using Parameter-elevation Regressions on the Independent Slopes Model (PRISM) for five Texas county locations, which included Deaf Smith, Ochiltree, Hansford, Moore, and Parmer, where wheat production was observed, are presented in

Table 1. Summary of PRISM for the five Texas counties in 1981–2017.

Parameters	Units	Average	Maximum	Minimum	Standard Deviation
Temperature Min	°C	5.6	18.8	−6.4	8.6
Temperature Max	°C	21.6	34.0	9.0	8.6
Precipitation	mm	42.0	90.6	11.7	23.5

. Once a location was determined, the latitude and longitude were obtained and used to gather data from PRISM. Daily maximum and minimum temperature and precipitation values were collected, averaged for the daily average, and summed together to obtain temperature and precipitation monthly totals; this was repeated for overall yearly totals.

Table 2. Summary of Texas county data from 1980 to 2022.

County	Lat	Long	Avg Yield	Max Yield	Min Yield	Median Yield	Standard Deviation	Elevation (m)
Deaf Smith	34.9170	−102.5979	2185.7	3201.2	1129.8	2185.7	544.7	1234
Ochiltree	36.1119	−100.6665	1923.4	3295.3	894.4	1856.1	544.7	893
Hansford	36.3386	−101.4570	2279.8	3328.9	1412.3	2138.6	524.6	982
Moore	35.9988	−102.1102	2656.4	3927.5	1681.3	2683.3	517.8	1140
Parmer	34.6710	−102.4811	2273.1	3147.4	1170.2	2259.6	504.4	1188
Average	-	-	2266.4	3382.7	1257.6	2226.0	-	-

Yields are measured in kg/ha.

displays each county’s latitude, longitude, average yield, maximum, and minimum yield, standard deviation, and elevation of wheat from 1980 to 2022. The selection of each county’s latitude and longitude was conducted through the evaluation of the Crop Data Layer (CDL) website. The accuracy of the CDL data on winter wheat-producing fields allowed the determination of latitude and longitude to be made within the middle of a large winter wheat-growing plot. Winter wheat-growing areas were analyzed in each of the five counties to determine the best location to gather coordinates to be used in the PRISM, CMIP5, and soil data collection.

The maximum yield was reached in 1992 with 3927.5 kg/ha in Moore County and the minimum of 894.4 kg/ha in Ochiltree County in 1995. The five counties had an overall average of 2266.4 kg/ha between 1980 and 2022.

The USDA-NRCS Soil Survey Geographic Database (SSURGO) for Texas has been assembled for the conterminous United States. For this study, the SSURGO data layer was overlaid on the Cropland Data Layer (CDL) data in order to locate longitude and latitude coordinates to help determine the soil types applicable to wheat fields in Texas. Crop management data were also obtained to determine the field operations for Texas winter wheat.

Table 3. Summary of wheat production in Texas from 1968 to 2020.

Parameters	Units	Average	Maximum	Minimum	Standard Deviation
Planted Acreage	mil ha	6.0	3.3	1.4	0.4
Yield	kg/ha	1896.5	2488.3	1076	309.4
Production	bil kg	2.6	5.1	0.8	0.9
Price Received	$/kg	0.17	0.28	0.08	0.06

summarizes the production of wheat in Texas from 1968 to 2020. The average, maximum, minimum, and standard deviation for winter wheat in Texas are displayed in the table. Although the yearly yield averages have increased in later years, the average displayed takes the lower yield totals from the earlier years into consideration. Improved seed genetics and management strategies have helped improve wheat production in Texas.

Table 4. Summary of the amount of fertilizer applied (lbs./acre) on Texas wheat from 1964 to 2017.

Fertilizer	Unit	Average	Maximum	Minimum	Standard Deviation
Nitrogen	kg/ha	88.3	125.6	52.7	37.2
Phosphate	kg/ha	43.8	60.5	25.8	19.7
Potash	kg/ha	23.9	54.9	4.5	14.3

is a summary of the fertilizers (nitrogen, phosphate, and potash) applied to Texas wheat production from 1964 to 2017. The nitrogen maximum was 125.6 kg/ha in 1978 and the minimum was 52.7 kg/ha in 2009. The phosphate maximum was 60.5 kg/ha in 1969 and the minimum was 25.8 kg/ha in 2017. The potash maximum was 54.9 kg/ha in 1987 and the minimum was 4.5 kg/ha in 1964.

2.2. APEX

APEX is a comprehensive field-scale model developed to evaluate the effects of management strategies on various agronomic and ecological indicators, such as crop growth and yields [22]. APEX is an updated version of the Erosion Productivity Impact Calculator (EPIC) model, which has been used to simulate different management scenarios, such as changes in manure and fertilizer application rates, mode and timing, alternative tillage practices, structural controls, and other cultural management practices [22,29]. The APEX model also has the capability to evaluate the effectiveness of conservation practices [36,37,38]. Figure 1 illustrates the inputs for the APEX model used in this study.

APEX runs on a daily time step and can be useful for a wide range of soil, crop rotation, climate, landscape, and management practice combinations. Due to its high detail, APEX simulates accurate management practices [30]. The model allows input of simulated cropping systems, fertilizer and/or manure nutrient characteristics, soil layer properties, tillage practices, and other characteristics for each subarea. A key model output is crop yield [29,37,38].

APEX was calibrated against annual county-level crop yield data that are available through the USDA National Agricultural Statistics Service (USDA-NASS), and historical weather, soil, and agricultural management strategies in Texas. For this study, APEX was validated against the available Texas winter wheat crop yield data collected from USDA-NASS. APEX was then used to estimate Texas winter wheat crop yields under future climate patterns from the different Representative Concentration Pathways (RCPs), which represent a trajectory of Green House Gas (GHG) emission concentrations.

The field operations that were entered into the APEX model are presented in

Table 5. Field operations simulated for Texas winter wheat, 1981–2017 production years.

Date	Operation
15 February	Apply nitrogen, phosphate, and potash
1 May	Harvest
30 June	Kill
15 July	Till
1 August	Disk
15 October	Plant
30 November	Apply nitrogen, phosphate, and potash

. The fertilizer rates were adjusted accordingly for each year to match the actual data gathered from Texas AgriLife fertilizer data. The planting and harvesting dates used were selected from the Texas AgriLife database. Management strategies were also an input in the file for the APEX model.

2.3. CMIP5

Climate projections from CMIP5 consider a broader range of options that include a Representative Concentration Pathway (RCP), which represents a trajectory of Green House Gas (GHG) emission concentrations expected in response to the corresponding mitigation assumptions [29]. This includes a lower, least intense, scenario than had been considered before RCP2.6. On the extreme end, the scenarios include one that assumes continued increases in emissions (RCP8.5) and the corresponding greater amount of warming. Also shown are temperature changes for the intermediate scenarios RCP4.5 and RCP6.0. Projections show a change in average temperature in the later part of this century from 2071 to 2099 relative to the later part of the last century from 1970 to 1999 [39]. Each climate change scenario used in this study is based on CMIP5 model projections. RCP2.6, RCP4.5, RCP6.0, and RCP8.5 are the four main RCPs adopted by the Intergovernmental Panel on Climate Change (IPCC) [29,40]. These four have been broadly used for climate projections. The most optimistic of the four for reducing global warming is RCP2.6 [41], which is projected to result in an increase of 1 °C in mean global temperatures by the time period 2046–2065 [29]. On the other hand, the most pessimistic scenario of the four for reducing global warming is RCP8.5 [29]. It is important to analyze the minimum and maximum temperature and precipitation patterns from the climate scenarios because the patterns are the foundation of the simulated or actual yield impacts of climate change. In this study, the climate projection data included a daily time series on minimum and maximum temperature and precipitation that were available for the time period of 2005 to 2099. For climate scenario simulations, the twelve scenarios used in this study are represented in

Table 6. List of different climate scenarios simulated.

Scenarios	1956–1980	1981–2005	1981–2017	2021–2045	2046–2070	2071–2095
Historical CMIP5	X	-	-	-	-	-
Historical PRISM	-	Baseline	-	-	-	-
PRISM	-	-	Baseline	-	-	-
RCP2.6	-	-	-	X	X	X
RCP4.5	-	-	-	X	X	X
RCP8.5	-	-	-	X	X	X

Each scenario-date combination marked “X” or “Baseline” was simulated for the five selected Texas counties.

. The twelve scenarios were conducted for each of the five counties, for a total of 60 scenarios.

3. Results

3.1. Data Analysis

3.1.1. Fertilizer Application in Wheat Production in Texas

In Figure 2, the fertilizer and rates applied in kg/ha and yields in kg/ha are shown from 1980 to 2017. The fertilizers and yields are all correlated throughout the 1980–2017 timespan. When fertilizer input increases during a year, yields during the same year increase as well. The same is true when there is a decrease in fertilizer usage and yields decrease in the same year. In the later years, some data values are missing, but there is still a visible trend between nitrogen, phosphate, and potash rates applied to Texas winter wheat fields.

3.1.2. Average Wheat Yields in Texas

Yearly USDA NASS average yields from the five Texas counties from 1980 to 2022 (Table 2) are displayed in Figure 3. In 2013, the lowest yield average can be observed at 1558 kg/ha; the next lowest is in 2011, with a yield average of 1665.1 kg/ha, which correlates to the drought experienced during the corresponding year. The maximum yield was reached in 1999 with 3191.7 kg/ha, followed by a significant drop in the average yield the following year (2000) to 1837.3 kg/ha.

3.2. Analysis of Historical Climate Patterns

The next figures represent the average monthly temperature (measured in °C), minimum, maximum, and precipitation for the five counties located in the top wheat-producing areas. Figure 4 illustrates the PRISM temperature and monthly minimum averages collected from 1981 to 2017 and is measured in °C. The graph gives a visualization of how the average monthly minimum decreases starting in November through March, with the coldest monthly average in January at −5.9 °C and an overall average temperature of 5.6 °C. Figure 5 illustrates the average monthly maximum temperatures for the five wheat-producing counties in Texas from the PRISM data between 1981 and 2017. The maximum temperatures were highest from June to September, with the highest at 33.4 °C in July and an overall yearly average of 21.6 °C.

Figure 6 presents the monthly average precipitation, measured in mm, from the same five Texas counties in the previous two graphs from 1981 to 2017. During the months of December to February, the least amount of precipitation was recorded in the PRISM data. The lowest amount of precipitation recorded in the PRISM data was in February with 13.7 mm, followed by a maximum of 77.2 mm in June. The yearly total average was 503.5 mm. In Figure 7, the monthly averages of all PRISM data are displayed from 1981 to 2017. Average temperature minimum (Avg TMN) and average temperature maximum (Avg TMX), and average precipitation (Avg PRCP) were measured in °C and mm, respectively.

Figure 8 is a visual representation of the temperature changes from the most recent CMIP5 model projections, which include the lowest scenario, RCP2.6, the highest scenario, RCP8.5, and two intermediate scenarios, RCP4.5 and RCP6.0. Projections in this figure show changes in the average temperature in the United States. It should be noted that, as temperatures rise in each of the CMIP5 scenarios, there is also an increase in the amount of precipitation. In the RCP2.6 scenario in Texas, temperatures are expected to increase by up to 2.2 °C. Under the RCP4.5 scenario, temperatures are expected to increase between 1.7 and 3.3 °C. In the RCP6.0 scenario, the temperature of the majority of Texas is expected to increase by up to 3.3 °C. Lastly, the most extreme scenario, RCP8.5, has projections for temperature increases between 3.9 and 5.6 °C, but mainly increases between 4.5 and 5 °C (

Table 7. CMIP5 scenario temperature increase comparison.

	United States Maximum	United States Minimum	Texas Maximum	Texas Minimum
RCP 2.6	3.3	<1.7	2.2	<1.7
RCP 4.5	5.6	1.7	3.3	1.7
RCP 6.0	>8.3	2.2	3.3	2.2
RCP 8.5	>8.3	3.3	5.6	3.9

Temperature in °C.

).

3.3. APEX and Simulated Impacts

Simulated Winter Wheat Yields under RCP Scenarios

Results from the APEX model created in this study are illustrated in Figure 9, Figure 10, Figure 11 and Figure 12. The simulations from this model between 1981 and 2095 were created with inputs from management practices, soil, PRISM, and fertilizer data for Texas wheat. Simulations show a decrease under each of the CMIP5 scenarios. Figure 9 displays the average yields from Deaf Smith, Ochiltree, Hansford, Moore, and Parmer counties. The highest average yield simulated compared to the validated baseline of PRISM 1981–2017 was in the RCP8.5 2071–2095 scenario with 2715.3 kg/ha, which indicates a 719.6 kg/ha yield increase. The minimum average yield result was 1531.7 kg/ha under the RCP4.5 2046–2070 scenario, which was a decrease of 464 kg/ha.

Table 8. PRISM and historical yields.

County	PRISM 1981–2017	Historical CMIP5 1956–1980	Historical PRISM 1981–2005
Deaf Smith	2091.5	1741.8	1627.5
Ochiltree	2205.8	1856.1	1970.5
Hansford	1856.1	1856.1	1741.8
Moore	1741.8	1627.5	1506.4
Parmer	2091.5	1856.1	1741.8
Average	1997.4	1788.9	1714.9

Yields are in kg/ha.

displays the yield results from the PRISM 1981–2017, Historical CMIP5 1956–1980, and Historical PRISM 1981–2005 scenarios. Results from these scenarios were validated against county-level data and accurately reflected yield results from Deaf Smith, Ochiltree, Hansford, Moore, and Parmer counties. Calibrating the model and validating results was crucial for the accurate projections conducted in this study.

Table 9. CMIP5 RCP2.6 yield simulations.

County	RCP2.6 2021–2045	RCP2.6 2046–2070	RCP2.6 2071–2095
Deaf Smith	1972.6	2204.7	2088.7
Ochiltree	2552.8	2784.9	2668.9
Hansford	2436.8	2552.8	2320.8
Moore	1972.6	2088.7	1972.6
Parmer	2088.7	2320.8	2204.7
Average	2204.7	2390.4	2251.1
Max	2555.5	2784.2	2669.9
Min	1970.5	2091.5	1970.5

Yields are in kg/ha.

and Figure 10 represent the CMIP5 RCP2.6 simulated yields for each county where temperatures are projected to increase up to 2.2 °C. The five counties all indicated yield increases in the RCP2.6 2046–2070 scenario when compared to the validated PRISM 1981–2017 data. The minimum yield from the RCP2.6 2021–2045 scenario was from Deaf Smith and Moore County, with a simulated yield of 1972.6 kg/ha. In the RCP2.6 2046–2075 and 2071–2095 scenarios, Moore County had the minimum simulated yield of 2088.7 kg/ha and 1972.3 kg/ha, respectively.

Table 10. CMIP5 RCP4.5 simulated yields.

County	RCP4.5 2021–2045	RCP4.5 2046–2070	RCP4.5 2071–2095
Deaf Smith	1508.5	1392.5	1972.6
Ochiltree	2204.7	1972.6	2784.9
Hansford	1972.6	1624.5	2204.7
Moore	1508.5	1276.4	1856.6
Parmer	1624.5	1392.5	2088.7
Average	1763.8	1531.7	2181.5
Max	2205.8	1970.5	2784.2
Min	1506.4	1277.8	1856.1

Yields are in kg/ha.

and Figure 11 represent the results of the simulated yields (kg/ha) under the CMIP5 RCP4.5 scenarios for 2021–2045, 2046–2070, and 2071–2095, where temperatures are projected to increase between 2.2 and 3.3 °C. Under the CMIP5 RCP4.5 scenario, the five calibrated counties all experienced yield increases in the 2071–2095 time period and decreases during the 2046–2070 time period when compared to the validated PRISM 1981–2017 data. Deaf Smith and Moore counties have a minimum simulated yield of 1508.5 kg/ha in the 2021–2045 time period. Moore County had the minimum simulated yields of 1276.4 kg/ha and 1856.6 kg/ha in the 2046–2070 and 2071–2095 time periods, respectively. The maximum simulated yields for the three time periods (2021–2045, 2046–2070, and 2071–2095) in the CMIP5 RCP4.5 scenarios all came from Ochiltree County with 2204.7 kg/ha, 1972.6 kg/ha, and 2784.9 kg/ha. The overall highest average yield increase for the five counties was evaluated in the 2071–2095 scenario, where temperatures are expected to increase by 3.3 °C and have a simulated average yield of 2181.5 kg/ha.

Table 11. CMIP5 RCP8.5 yields.

County	RCP8.5 2021–2045	RCP8.5 2046–2070	RCP8.5 2071–2095
Deaf Smith	1972.6	2204.7	2204.7
Ochiltree	2204.7	3017.0	4061.3
Hansford	2088.7	2552.8	3017.0
Moore	1740.6	2204.7	2088.7
Parmer	2088.7	2320.8	2204.7
Average	2019.1	2460.0	2715.3
Max	2205.8	3019.6	4062.0
Min	1741.8	2205.8	2091.5

Yields are in kg/ha.

and Figure 12 represent the calibrated yield results from the CMIP5 RCP8.5 simulated projections from 2021–2095 in 25-year increments. In the RCP8.5 scenarios, Texas areas are projected to see temperature increases between 3.9 and 5.6 °C, but the majority will see increases between 4.5 and 5 °C. Four of the five counties experienced yield increases under the RCP8.5 2045–2070 and 2071–2095 scenarios when compared to the PRISM 1981–2017 validated results. Deaf Smith was the only county out of the five that experienced a 5.6% decrease in the RCP8.5 2021–2045 projections. Moore County had the minimum simulated yields of 1740.6 kg/ha and 2088.7 kg/ha during the 2021–2045 and 2071–2095 scenarios, respectively. Under the 2046–2070 time period scenario, Deaf Smith and Moore counties both had the minimum simulated yield of 2204.7 kg/ha. Of the five counties, Ochiltree County had the maximum simulated yield from all three time periods (2021–2045, 2046–2070, and 2071–2095) with 2204.7 kg/ha, 3017 kg/ha, and 4061.3 kg/ha, respectively. The CMIP5 RCP8.5 scenario is the most intense, with the highest temperature and precipitation increases. Under this scenario, the five counties still experienced yield increases compared to the PRISM 1981–2017 validated results. This occurrence is primarily due to increases in precipitation and the northern part of Texas experiencing a more favorable wheat-growing environment.

Table 12. Percentage changes in winter wheat yields under climate scenarios relative to values under PRISM weather 1981–2017 results (PRSIM 1981–2017, Historical CMIP5 1956–1980, and Historical PRISM 1981–2005. RCPs are 2021–2045, 2046–2070, and 2071–2095).

	Historical			RCP 2.6			RCP 4.5			RCP 8.5
	81–17	56–80	81–05	21–45	46–70	71–95	21–45	46–70	71–95	21–45	46–70	71–95
Deaf Smith	0.0	−16.7	−22.2	−5.6	5.6	0.0	−27.8	−33.3	−5.6	−5.6	5.6	5.6
Ochiltree	0.0	−15.8	−10.5	15.8	26.3	21.1	0.0	−10.5	26.3	0.0	36.8	84.2
Hansford	0.0	0.0	−6.3	31.3	37.5	25.0	6.3	−12.5	18.8	12.5	37.5	62.5
Moore	0.0	−6.7	−13.3	13.3	20.0	13.3	−13.3	−26.7	6.7	0.0	26.7	20.0
Parmer	0.0	−11.1	−16.7	0.0	11.1	5.6	−22.2	−33.3	0.0	0.0	11.1	5.6

The baseline is PRISM 1981–2017.

displays the percentage changes of winter wheat yields under climate scenarios relative to values under PRISM weather. The baseline for Table 12 is the PRISM 1981–2017 validated results. Looking at the five calibrated counties, Deaf Smith and Parmer counties experienced the greatest yield decrease of 33.3% under the CMIP5 RCP4.5 (Texas projected temperature increase between 3 and 6 ℉) 2046 to 2070 scenario compared to the baseline. The maximum percentage yield increase was noticed in Ochiltree County under the CMIP5 RCP8.5 (Texas projected temperature increase between 7 and 10 ℉) 2071 to 2095 scenario, with an 84.2% yield increase compared to the baseline. Fourteen out of the 15 RCP8.5 scenarios simulated for the five calibrated counties displayed either no change or an increase in yields, except for Deaf Smith, which experienced a 5.6% decrease in yields in the RCP8.5 2021–2045 time period.

In general, to provide a realistic assessment of climate change impacts, it is important to augment the yield impacts presented above with the long-term yield trajectory. This is because the climate change impacts are really displacements from a long-term yield trajectory of the crop that reflects genetic improvements and management practice changes. Consequently, the actual yields anticipated would be the sum of the long-term yield trend effect and the climate change effect. However, in this specific study, the historical yield trend for winter wheat in the selected counties—the combined result of agronomic and genetic (breeding) factors—was found to be basically flat for the past 43 years (see Figure 3) based on data from USDA-NASS QuickStats. Consequently, the yield projections are presented here without adjusting for a long-term yield trend, as the latter practically had a zero slope. We point out that the relatively flat winter wheat yield trajectory for the study area in Texas is not necessarily reflective of the winter wheat yield trend for the entire U.S.

Table 13. Percentage changes in winter wheat yields under climate scenarios relative to values under Historical PRISM weather 1981–2005 results. PRSIM 1981–2017, Historical CMIP5 1956–1980, and Historical PRISM 1981–2005. RCP scenario time frames are 2021–2045, 2046–2070, and 2071–2095.

	Historical			RCP 2.6			RCP 4.5			RCP 8.5
	81–17	56–80	81–05	21–45	46–70	71–95	21–45	46–70	71–95	21–45	46–70	71–95
Deaf Smith	28.6	7.1	0.0	21.4	35.7	28.6	−7.1	−14.3	21.4	21.4	35.7	35.7
Ochiltree	11.8	−5.9	0.0	29.4	41.2	35.3	11.8	0.0	41.2	11.8	52.9	105.9
Hansford	6.7	6.7	0.0	40.0	46.7	33.3	13.3	−6.7	26.7	20.0	46.7	73.3
Moore	15.4	7.7	0.0	30.8	38.5	30.8	0.0	−15.4	23.1	15.4	46.2	38.5
Parmer	20.0	6.7	0.0	20.0	33.3	26.7	−6.7	−20.0	20.0	20.0	33.3	26.7

The baseline is Historical PRISM 1981–2005.

shows the percentage changes in winter wheat yields under climate scenarios relative to values under Historical PRISM weather 1981–2005. Looking at the five calibrated counties, Parmer County experienced the greatest yield decrease of 20% under the RCP4.5 2046–2070 scenario when compared to the baseline. The maximum percentage simulated yield increase occurred in Ochiltree County, with a 105.9% increase under the RCP8.5 2071–2095 scenario when compared to the baseline. The five calibrated counties all experienced yield increases when compared to the baseline in all of the RCP2.6 and RCP8.5 scenarios.

4. Conclusions and Recommendations

4.1. Conclusions

The purpose of this study was to contribute to the understanding of the projected changes in climate by estimating the impacts on winter wheat yields in Texas. Yield results from the APEX model were validated against actual Texas winter wheat yield data from USDA NASS. The production and prices of wheat in Texas between 1968 and 2020 have given insight into the effects of factors such as weather, management strategies, and location on the total outcome. This study has shown trends and changes over time and provides valuable knowledge to determine future yield rates and pricing related to winter wheat. Through the analysis of historical production and price data, climate change, such as increases in temperature and droughts, has had some impact on Texas winter wheat but not as drastically as projected due to precautionary measures taken. This study will contribute knowledge to aid farmers’ understanding of the impacts of climate variability on winter wheat yields and prepare for the impacts if climate change occurs.

Texas winter wheat yields in this paper are projected to increase in the future by utilizing the most broadly used projected climate change scenarios. Through the use of a calibrated APEX model, Deaf Smith and Parmer counties experienced the greatest yield decrease of 33.3% under the CMIP5 RCP4.5 2046–2070 scenario, and a maximum percentage yield increase was noticed in Ochiltree County under the CMIP5 RCP8.5 2071–2095 scenario with an 84.2% yield increase compared to the PRISM 1981–2017 baseline. Parmer County experienced the greatest yield decrease of 20% under the RCP4.5 2046–2070 scenario and a maximum percentage simulated yield increase occurred in Ochiltree County with a 105.9% increase under the RCP8.5 2071–2095 scenario, when compared to the Historical PRISM 1981–2005 baseline. Even with projected temperatures increasing, the five calibrated counties exhibited increases in yields. This is a result of the increased precipitation that accompanies the projected rising temperatures. In northern Texas counties, these weather changes will make wheat production more suitable in the future. Although the impacts of extreme temperatures and fluctuations in precipitation are expected to have a negative impact on yields of winter wheat in Texas, some of the yield decreases are not certain.

4.2. Recommendations

Improvements to farming strategies and management will help reduce the projected yield decreases. Enhancements to seed genetics, irrigation, and crop migration are all positive factors that can help offset the negative impacts of the projected climate change effects on winter wheat yields in Texas. Due to these advancements and improvements, Texas winter wheat yields may be able to maintain average yield values or even slightly improve in the future under climate change trends. Some areas in Texas that were once considered too cold will become more suitable for winter wheat crops and help maintain yields. This can potentially offset the projected yield decreases based on the five counties used in this study. Overall, yields will not be as heavily impacted as projected. Lastly, it is difficult to predict the future strategies and inputs of producers and the overall impact they will have on total crop yields. Studies will continue to analyze the impacts of climate variability on winter wheat yields in Texas to help farmers form strategies to combat the potential detrimental effects and preserve global food security.