Assessment of Corn Grain Production Under Drought Conditions in Eastern Mexico Through the North American Drought Monitor

Abstract

Over 80% of corn on Mexico’s eastern side is sown under rainfed conditions. Therefore, drought represents a constant challenge for local producers. This study aims to determine the effects of drought on rainfed corn grain production on Mexico’s eastern side by using the North American Drought Monitor as the primary tool. Drought levels at the municipal level provided by this monitor and corn production data (surface damage, yield, and volume) of the two productive seasons (spring–summer and autumn–winter) during 20 years were correlated at two significant levels (0.05 and 0.01). The significant values (p < 0.05) were used to obtain regression curves representing corn-drought behaviors. The National Disaster Statistics and climatological stations were considered, discarding other phenomena besides drought. Results indicate that, for the significant municipalities, the years with the highest drought levels (2005, 2011, and 2019) positively correlate with reduced corn grain yield, volume, and total harvest losses. The regression curves estimated a yield reduction of 78 kg∙ha⁻¹ during the spring–summer season and 76 kg∙ha⁻¹ during the autumn–winter season. We concluded that the Drought Monitor is valuable for determining relationships between rainfed corn grain productivity and drought, considering that no other climatological phenomena affect the region.

1. Introduction

1.1. The North American Drought Monitor

Drought is a low-evolution phenomenon that is challenging to predict due to its hazardous nature. It is characterized by abnormally dry conditions that cause a hydrological imbalance that can have destructive effects [1]. According to the impact drought causes, it can be classified as meteorological, agronomical, hydrological, and socioeconomic [2].

Since drought is considered an extreme event, the United States of America, Canada, and Mexico implemented a cooperative effort among their drought experts to continuously monitor drought across their regions. This cooperation gave birth to the North American Drought Monitor [3]. In Mexico, the National Meteorological Service named its side the Mexican Drought Monitor [4]. According to the North American Drought Monitor, drought levels are established by considering local monitors of air, fires, precipitations, soil moisture, streamflows, temperatures, vegetation indexes, and dams’ water percentage [3]. Satellite images are also used to compose the tool and make the information available. This tool provides historical statistics and a map showing where drought occurs, its spatial extent, its severity, and the time scale of the associated impacts. It also includes information about the different drought levels in each municipality of the Mexican territory. According to this monitor, six possible levels represent drought conditions: (1) 0—No drought conditions were determined in the region. (2) D0—Abnormally dry: when it occurs at the beginning of a drought, short-term dryness may cause slow planting and growth of crops or pastures. When it occurs at the end of a drought period, it may cause lingering water deficits and pastures or crops not fully recovered (3) D1—Moderate drought: which may cause several crop and pasture damages, a high risk of fires, and low levels in rivers, streams, water reservoirs, and wells. (4) D2—Severe drought: associated with a high probability of crop damages, a high risk of fires, water shortage, and limits in water consumption. (5) D3—Extreme drought: indicates high crop or pasture losses, extreme risk of forest fires, and a general restriction in water consumption. (6) D4—Exceptional drought: indicates widespread crop or pasture losses and shortages of water reservoirs, streams, and wells, creating water emergencies [3]. With these levels, the Drought Monitor provides 15-day reports for each municipality of the Mexican territory, which can be used by decision-makers, producers, and ordinary citizens interested in the evolution of drought for research purposes [4]. Considering these definitions, it can be assumed that meteorological, agricultural, and hydrological droughts are the primary drought types for the Drought Monitor [5].

1.2. Historical Drought Levels in Mexico and Their Relationship with Agricultural Data

Mexico is a country where drought usually affects more than 50% of its territory [6]. Nowadays, this situation worsens if we consider that drought periods are increasing in more extensive parts of the country. For example, according to the last report of the Weather Meteorological Organization in Latin America and El Caribe, drought and high temperatures in 2022 have increased in average values [7].

The historical records of the colonial period and the modern Mexican Drought Monitor [4] indicate that Mexico has experienced periodical droughts [8]. Nevertheless, the Drought Monitor shows that there have been more increments in their intensity since 2011, when more than 85% of the territory recorded D1 or above levels of drought. Regarding agricultural data, old records indicate that severe drought has caused the loss of multiple crops in the southeastern side of Mexico (mainly in the Yucatan peninsula), forcing people to displace and leave their home cities since pre-colonial and colonial times [9]. More recently, three significant dry periods were registered in the Mexican territory for the last century: 1948–1954, 1970–1978, and 1993–1998. During the first period, the northeastern side of Mexico (above the parallel 22°) reported 250,000 ha of cotton lost by the drought. The eastern side of Mexico was entirely affected during the second period, causing the total loss of corn and cotton sown. For the last period, all the territory was affected by drought. During this period, corn and wheat were reported to have the highest damaged surfaces, with more than 300,000 ha lost. Although this information does not mention if the affected surfaces were rainfed or irrigated lands, these reports specify that corn cultivation generally recorded the highest surface damage during drought in eastern Mexico [2].

1.3. Corn Production in Mexico

Corn (Zea mays L.) is the most cultivated crop in Mexico, sown in every state of the Mexican Republic [10,11]. It can be planted under different agroclimatic conditions, from sea level up to 3000 m above sea level and in wet to dry climates [12]. Therefore, corn cultivation is subject to many hydrometeorological phenomena, like excessive rains, strong winds, cold temperatures, and droughts. In the country, corn has different applications: corn grain, which is sold as cob grain; fodder corn, which is sold as pasture for animal feed; popcorn corn, a unique variety for human consumption; and corn seed, sold for sowing purposes [11]. From these varieties, corn grain is the most cultivated variety in the country, with more than 27.50 tons harvested in 2022 [13]. It is also used for animal and human consumption, and the yellow and white varieties are the most common sown in the country.

1.4. Corn Grain on the Eastern Side of the Mexican Territory

On the eastern side of Mexico, along the Gulf of Mexico, most of the territory includes two states: Veracruz in the south and Tamaulipas in the north. This area combines a mix of tropical, subtropical, semidry, and dry climates, where agricultural activities are an essential part of living for many local producers. Despite climate conditions, seasonal rainfed crops, which are sown during specific periods of the year and depending on local climatic conditions, are sown yearly, regardless of possible drought conditions. In Mexico, according to the National Service for Agrifood and Fisheries Information (SIAP), there are two seasons for sowing and harvesting: spring–summer and autumn–winter. The spring–summer period lasts from April to September, while the autumn–winter period lasts from October to March. In this context, corn is sown in most Mexican territories in both seasons [11].

In both states (Veracruz and Tamaulipas), corn grain is sown under seasonal conditions in 247 municipalities out of their 255 locations. This situation represents a yearly average of 444,861 hectares of land under rainfed corn grain cultivation [11]. Previous research on the performance of different crops correlated with climatic phenomena in part of the eastern side of Mexico indicates that tropical cyclones and floods have the highest number of significant correlations with damaged cornland. However, this study also shows that drought is in third place, causing total rainfed corn grain losses of 63,705 ha during the most recent worst drought periods registered on this site [14]. In other studies on Mexico’s western side, corn has also been affected by prolonged droughts, with a 43% reduction in crop grain productivity during the dry seasons [15]. One local study in the east determined a positive correlation between drought and corn productivity; however, the area under study was only a tiny fraction of the eastern central side of Mexico (8605 km²) [16]. Besides this, according to future climate change scenarios [17], there will be an increase in drought periods and temperatures in the Mexican territory.

Since corn is an essential food of the Mexican diet, its production and its relationship with climate data require an analysis where it is necessary to access local climatologic records and corn productivity per site. In this regard, the National Meteorological Service of Mexico (SMN) is the governmental entity responsible for maintaining climatic records from the Mexican territory [18]. The SMN contains historical data since the beginning of each local station installation. However, the eastern side of Mexico has poor coverage, with only approximately 31% of the information accurately updated [19]. This situation is critical to determining climatic conditions because this region has a complex geography. First, the whole area has more than ten longitudinal degrees of length, and second, mountain chains cross from south to north along the territory, with some sites at sea level and, in contrast, others up to 4000 m above sea level (masl) in less than 200 km of distance from east to west [20,21].

Due to this lack of information, one study aimed to predict agricultural drought behavior in the short term for the central part of the Gulf of Mexico using a Long Short-Term Memory Network and Landsat 8–9 multispectral images. It required the analysis of 183 high-resolution images for an area of only 5177.9 km² for 10 years. However, the prediction level was low when high and low droughts occurred in regions a few kilometers away [22]. These prediction difficulties can be explained by another study suggesting that 30 or more samples should be used to feed a model [23]. In this case, this model additionally would need to consider different seasonal and agricultural periods. For example, when the land is uncovered because of the tillage, when crops cover it completely, and the natural seasons of the year. These different land covers increase the difficulties in detecting drought levels in territories dedicated to agronomical activities.

Therefore, considering the extensive territory of the eastern Mexican side and the difficulties of analyzing drought indexes in municipalities without climatological information, the drought levels provided by the Drought Monitor are the only ones available for all the municipalities cultivating corn grain. The Drought Monitor is also considered the most holistic measure of drought [5]. Besides this situation, corn grain productivity and its relationship with drought levels, like the ones reported by the North American Drought Monitor, have not been studied yet in a land of 152,073 km², which is the one that comprises Veracruz and Tamaulipas. For example, the North American Drought Monitor periodically reports drought conditions in cornfields in the United States of America [24], which is absent on the Mexican side.

In addition, there is no information regarding drought levels occurring during governmental Declarations of Disaster (DD) by drought. A DD by drought is issued when drought levels cause crop or cattle losses or when the habitant’s capacity to obtain water for their primary and domestic uses is compromised [25]. Therefore, learning the link between drought levels and corn grain reduction and losses will allow producers to evaluate historical corn development under arid conditions, especially considering climate change.

This research aims to evaluate corn grain productivity under different drought levels provided by the Mexican Drought Monitor on the eastern side of Mexico (Veracruz and Tamaulipas states).

2. Materials and Methods

2.1. Research Site and Analysis Period

The total territory is situated at 27°40′45″ north latitude, 100°08′42″ west longitude, 17°10′00″ south latitude, and 93°36′29″ east longitude [20,21]. On the southern side of the territory (below the parallel 23°), the more considerable extension has a warm sub-humid climate, with summer rains, in 54.13% of the territory, while on the northern side of the territory, a semiwarm subhumid climate, with summer rains, occupies 29.38% of the territory, followed by a semidry warm climate, with summer rains, in 23.29% of the territory. Other climate types, like temperate humid, warm dry, temperate dry, and semicold subhumid, comprise 3.2% of the territory.

Since the agrifood system databases and the Drought Monitor used for this research contain historical information at the municipal level starting in 2003, this research considers only data from 2003 to 2022 (20 years), the available information period for both databases.

2.2. Extreme Climatic Conditions Causing Disaster Declarations

Besides the Drought Monitor, extreme climatic conditions were considered in this study because they may affect crop fields during their occurrence, and, in these cases, the Mexican government issues a Disaster Declaration at a municipal level. The National Center for Disaster Prevention (CENAPRED) is the official agency that maintains a database with DD containing the dates and climatic phenomena causing the catastrophe [25]. Therefore, data can be obtained from CENAPRED and compared with corn production to determine other phenomena besides drought affecting the crop.

Table 1. Classification of climatic phenomena causing disaster declarations in Mexico.

Identification	Climatic Conditions
Tropical cyclones	They can be tropical storms, tropical cyclones, and hurricanes.
Rains	Heavy or excessive precipitation that was not expected and caused severe losses.
Floods	Floods occur when a river, lake, or lagoon overflows. They can also happen when a specific place is flooded, causing losses to properties, people, cattle, or crops.
Low temperatures, snow, frost, or hail	They can include low temperatures, snow, frost, or hail, causing damage to people or crops.
Winds	These are strong winds destroying properties or crops.
High temperatures	They are extreme temperatures causing health problems for humans and animals.
Droughts	Droughts are considered when rains do not occur during their typical period. This phenomenon causes low water reservoirs, lack of water, and cattle and crop losses.

Source: National Center for Disaster Prevention [25].

defines the most common climatic phenomena registered by CENAPRED and the classification this research considers.

2.3. Meteorological Information

The meteorological information considered temperature and precipitation obtained from the national meteorological stations in the area [26]. This information was used to explain the historical climatological behavior of the sites where meteorological information is available. Of the 283 available meteorological stations operating in the area, only 38 contained the minimum 80% of data recommended by the World Meteorological Organization [27] for the 42-year base period (1980–2022). This extended period was considered because the base period reported by the Mexican Meteorological Service started in 1980 and concluded in 2010. However, significant drought levels have been recorded in recent years (2011, 2013, and 2019) [4], which were considered in this study.

Figure 1 shows the location of the available meteorological stations in the study region and its most representative climate types. The information on each station is shown in

Table A1. Geographical location of the meteorological stations used in this study.

Meteorological Station	Municipality/State	Latitude (°)	Longitude (°)	Altitude (msnm)
Juan Rodríguez Clara	Juan Rodríguez Clara, VER.	17.9930556	−95.402778	148
El Raudal	Nautla, VER.	20.1561111	−96.721111	10
Tecolutla	Tecolutla, VER.	20.4791667	−97.009722	7
Actopan	Actopan, VER.	19.5027778	−96.611111	250
Alamo	Álamo Temapache, VER.	20.9294444	−97.679444	19
Angel R. Cabada	Ángel R. Cabada, VER.	18.5972222	−95.447222	28
Villa Tejeda	Camarón de Tejeda, VER.	19.0222222	−96.613889	348
Chicontepec De Tejeda (SMN)	Chicontepec, VER.	20.9933333	−98.163889	291
Coatzacoalcos (OBS)	Coatzacoalcos, VER.	18.1402778	−94.522222	16
Rancho Viejo	Emiliano Zapata, VER.	19.4469444	−96.783611	914
El Tejar	Medellín de Bravo, VER.	19.0672222	−96.158333	10
Pánuco (DGE)	Pánuco, VER.	22.0591667	−98.175556	11
Loma fina	Paso de Ovejas, VER.	19.2613889	−96.41	41
Tempoal de Sánchez	Tempoal, VER.	21.5188889	−98.410278	34
Tuxpan (OBS)	Tuxpan de Rodríguez Cano, VER.	20.9597222	−97.418889	5
José Cardel	La Antigua, VER.	19.3647222	−96.374444	28
Platón Sánchez	Platón Sánchez, VER.	21.2980556	−98.356389	57
Poza Rica	Poza Rica de Hidalgo, VER.	20.5408333	−97.472778	50
San Juan Evangelista (DGE)	San Juan Evangelista, VER.	17.8833333	−95.145833	18
Ciudad Victoria (OBS)	Victoria, TAM.	23.7477778	−99.171667	336
Soto la Marina (OBS)	Soto la Marina, TAM.	23.7666667	−98.2	21
San Nicolás	San Nicolás, TAM.	24.6894444	−98.829722	797
Palmillas	Palmillas, TAM.	23.3022222	−99.548333	1264
Nuevo Morelos	Nuevo Morelos, TAM.	22.5283333	−99.213889	260
Miquihuana	Miquihuana, TAM.	23.5738889	−99.753056	1851
E.T.A. 067 CIUDAD MANTE	El Mante, TAM.	22.7425	−98.972222	91
Güémez	Güémez, TAM.	23.9186111	−99.004444	229
La Mayeb	González, TAM.	22.9155556	−98.345556	1371
Ahualulco	Gómez Farías, TAM.	22.9886111	−99.145	99
Cruillas	Cruillas, TAM.	24.7541667	−98.534722	229
Paso de Molina	Casas, TAM.	23.725	−98.745278	150
Bustamante	Bustamante, TAM.	23.4358333	−99.754167	1666
El Carrizal	Aldama, TAM.	22.8586111	−98.233333	90
San Gabriel	Xicoténcatl, TAM.	23.0841667	−98.7875	135
Conrado Castillo	Villagrán, TAM.	24.6669444	−99.253889	309
Jaumave (DGE)	Jaumave, TAM.	23.4075	−99.375278	324
El Barranco	Altamira, TAM.	22.5658333	−97.905	6
Abasolo (DGE)	Abasolo, TAM.	24.0655556	−98.39	70

.

2.4. Corn Productivity Data

Data for this research considered the 247 municipalities in Veracruz and Tamaulipas with historical records of corn grain production. Corn grain data were obtained from the Mexican National Service for Agrifood and Fisheries Information (SIAP) [11]. This service has provided municipal data since 2003. Data available consider the surface of land sown (ha), the surface damaged (although without information about the cause), the surface of land harvested (ha), volume harvested (tons), yield in tons by a hectare (tons∙ha⁻¹), and the selling price by ton (Mexican pesos tons⁻¹). This data is provided for each season (spring–summer and autumn–winter).

2.5. Drought Levels and Drought Intensity Estimations

Historical drought levels at the municipal level were obtained from the National Drought Monitor of Mexico, which started in January 2003. The drought levels provided by the Mexican Drought Monitor were converted to numerical values as follows: 0 = no drought condition, 1 = D0, 2 = D1, 3 = D2, 4 = D3, and 5 = D4, to quantify drought intensity.

Since corn grain production data considers the whole six-month period (spring–summer and autumn–winter seasons), equivalent drought levels were estimated for the same period according to Equation (1).

It_season = (ΣD0×1 + ΣD1×2 + ΣD2×3 + ΣD3×4 + ΣD4×5)/m

(1)

where It_season is the ponderate seasonal drought level of the municipality n.

ΣD0, ΣD1, ΣD2, ΣD3, and ΣD4 represent the sum of drought levels reported by the Drought Monitor at n municipality, and m is the number of reports during the seasonal period.

Additionally, the mean drought level by season per site period (the average drought level over the 20 years) was estimated to consider positive and negative oscillations.

Equation (2), which estimates the period’s summary drought intensity by each season (Itot_season), was used to obtain the municipalities with the highest drought records from 2003 to 2022.

Itot_season = ΣD0×1 + ΣD1×2 + ΣD2×3 + ΣD3×4 + ΣD4×5

(2)

The level of drought occurring during a Declaration of Disaster by drought was evaluated considering the maximum and average drought level recorded by the Drought Monitor during the period the statement was issued in the affected municipality.

2.6. Corn Yield and Volume Variation by Year and Season

Corn grain data per site summarized the number of hectares sown, harvested, or damaged by all the communities in the municipality. Corn grain yield was estimated as the average of all the municipalities (tons harvested over sown surface). Each period’s average values (surface sown, surface harvested, surface damaged, yield, and volume) per site were estimated to determine the total performance of the 2003–2022 period.

Linear regressions were estimated to determine the performance of yield and volume values during the analysis period. The regressions indicated that yield and volume values increased periodically during the study period. We assumed this increase could originate from new technologies or corn varieties. Therefore, for this study, the yield and volume per year and season were estimated using only the previous year of data, as shown in Equation (3).

yValue = current year value − previous year’s value

(3)

where yValue can be the yield or volume decrement or increment recorded during the current year, the current year value can be the yield or volume recorded during the current year, and the previous year’s value means the yield or volume of the prior year.

This way, we obtained the data series (yield and volume variation by year and season) to compare drought intensity by year and season.

2.7. Damaged Corn Surfaces and Drought Levels

Only years with drought declarations and the municipalities affected were considered in determining the correlation between damaged corn surfaces and drought levels because the study region registers many climatological phenomena other than drought. Drought levels and their intensities were considered using Equations (1) and (2) to correlate with damaged corn surfaces.

2.8. Summary of Sources of Information and Data Resolution

Table 2. Database sources for the information consulted during this research.

Data	Period	Spatial Resolution	Source
Climatological data	1980–2022	By local meteorological station	Mexican National Meteorological Service [26]
Land use	2020	1:250,000	Mexican Institute of Geography and Statistical Information [28]
Climate type	2001	1:250,000	Mexican Institute of Geography and Statistical Information [28]
Drought levels	2003–2022	By municipality	Mexican Drought Monitor is provided by the National Meteorological Service [4]
Disaster Declarations	2003–2022	By municipality	Mexican National Center of Disaster Prevention [25]
Corn grain productivity	2003–2022	By Locality	Mexican National Service for Agrifood and Fisheries Information [11]

summarizes the databases, period of study, and the spatial resolution of data used in this study for better clarification. The land use and climate types of cartographic information were considered to complement the discussion over the impacts of drought on corn grain productivity in relationship with their climate types and areas of agricultural land use.

2.9. Data Analysis

The Pearson correlation was used to determine the correlation coefficients (r) between drought levels and the number of declarations by drought and between crop data and drought levels. The statistical tests were performed using the software SigmaPlot 11.0.

For both seasons (spring–summer and autumn–winter), Disaster Declarations by drought, corn production (yield and volume), and its relationships with drought levels by municipality and year were correlated with a 0.05 and 0.01 significance. The Argis Desktop 10.8 software was used to show results with cartographic information. Since the damaged surfaces registered by the SIAP do not provide the reasons for the damage, only the municipalities with Disaster Declarations by drought provided by CENAPRED were considered to exclude other phenomena besides drought, as explained in Table 1.

Only data with significant correlations (p < 0.05) were selected to estimate the regression equations with the best fit (r²) to determine drought intensity in relation to corn grain yield and volume production and drought intensity versus damaged surface by using the SigmaPlot 11.0 program. Figure A1 shows the flow diagram of the analytical process.

3. Results

3.1. Drought Levels of the Study Site

Figure 2a shows the study region’s average drought levels per season. During the spring–summer season, the maximum average drought level was D1. During the autumn–winter season, two municipalities on the northeastern territory registered average D2 levels. However, only 9.0% of municipalities did not report drought during spring–summer, while 40.8% did not report drought during the autumn–winter. Figure 2b shows the summary of drought intensities of the municipalities during the period 2003–2022. Above parallel 25°, the northern side of the region has the highest drought levels during spring–summer, followed by the southern region, below parallel 18°. The highest sum of drought levels (351) on the northern side, above parallel 26°, indicates that 58% of the records were D3 and D4 drought levels.

Figure 2. Drought intensities in the study region. (a) Estimated average drought levels and their intensity in the study region during 2003–2022. (b) Sums of drought levels (Itot_season) recorded by municipality during 2003–2022. 0: no drought, D0: abnormally dry, D1: moderate drought, D2: severe drought. Sources: geographic information was obtained from the Mexican Institute of Geography and Statistical Information [29]. Drought data was obtained from the Mexican Drought Monitor [4]. Figure 3 shows the historical drought levels recorded in the study site. It can be seen that 2005, 2011, 2019, and 2020 were the years with the highest drought levels. The spring–summer season generally accounts for the highest drought levels (Figure 3a). For example, during spring–summer 2011, 47% of the municipalities recorded D3 drought levels, and 2% (six municipalities) recorded D4 levels, while during the autumn–winter season, only 4% of the sites recorded D3 levels, and D4 levels were not recorded (Figure 3b). During spring–summer 2019, 37% of the sites recorded D3 levels and 4% D4; during autumn–winter, only 6% recorded D3 levels, and D4 levels were not registered. By 2020, during the spring–summer season, from June to September, the number of municipalities without drought decreased from 93% to 35%, descending to 25% in April 2021 and 5% in May 2022.

Figure 2. Drought intensities in the study region. (a) Estimated average drought levels and their intensity in the study region during 2003–2022. (b) Sums of drought levels (Itot_season) recorded by municipality during 2003–2022. 0: no drought, D0: abnormally dry, D1: moderate drought, D2: severe drought. Sources: geographic information was obtained from the Mexican Institute of Geography and Statistical Information [29]. Drought data was obtained from the Mexican Drought Monitor [4]. Figure 3 shows the historical drought levels recorded in the study site. It can be seen that 2005, 2011, 2019, and 2020 were the years with the highest drought levels. The spring–summer season generally accounts for the highest drought levels (Figure 3a). For example, during spring–summer 2011, 47% of the municipalities recorded D3 drought levels, and 2% (six municipalities) recorded D4 levels, while during the autumn–winter season, only 4% of the sites recorded D3 levels, and D4 levels were not recorded (Figure 3b). During spring–summer 2019, 37% of the sites recorded D3 levels and 4% D4; during autumn–winter, only 6% recorded D3 levels, and D4 levels were not registered. By 2020, during the spring–summer season, from June to September, the number of municipalities without drought decreased from 93% to 35%, descending to 25% in April 2021 and 5% in May 2022.

Figure 3. Historical drought records in the study region during (a) spring–summer and (b) autumn–winter. 0: no drought, D0: abnormally dry, D1: moderate drought, D2: severe drought, D3: extreme drought, D4: exceptional drought. Source: drought data was obtained from the Mexican Drought Monitor [4].

Figure 3. Historical drought records in the study region during (a) spring–summer and (b) autumn–winter. 0: no drought, D0: abnormally dry, D1: moderate drought, D2: severe drought, D3: extreme drought, D4: exceptional drought. Source: drought data was obtained from the Mexican Drought Monitor [4].

3.2. Disaster Declarations and Drought Levels

Although there have been only five years with records of Disaster Declarations caused by drought, the levels of drought registered in the affected municipalities differed widely. Figure 4 shows the levels of drought registered during the 200 Disaster Declarations for the autumn–winter season and 132 recorded for the spring–summer season during 2003–2002. The lowest average level of drought recorded during Drought Declarations was D0, with only three municipalities indicating this level. In contrast, the highest average drought level recorded was D4, with nine sites recording it. The correlation coefficients between drought levels and Disaster Declarations by drought were significant at p < 0.01 for the autumn–winter season (r = 0.303) and the spring–summer season (r = 0.340).

Figure 5 shows the number of Disaster Declarations by drought in each municipality and its corresponding average drought level recorded during the declarations. The autumn–winter season registered 60% of the total number of Disaster Declarations of the two seasons, with 100% of the municipalities having at least one Disaster Declaration due to drought above the parallel 22.5° and two sites having up to three Disaster Declarations (Figure 5a). Nevertheless, the drought levels during the spring–summer season were higher than during the autumn–winter season, with average levels up to D2, D3, and D4 above parallel 25°. In the South, below parallel 18°, three sites reached D2 levels (Figure 5b). The highest number of Disaster Declarations by drought in two municipalities (Tantoyuca and Camarón de Tejeda (Figure 5a) was three during the autumn–winter season and two during the spring–summer season.

Figure 6 shows the monthly temperature behavior during four years with Disaster Declarations by drought: 2005, 2011, 2013, and 2019, in relation to the long-term climatology (1980–2022). During 2005 (Figure 6a). Positive anomalies occurred during the spring–summer (June—September) and January (autumn–winter season). These positive anomalies oscillated between 2% and 8% of the total range and represented temperature increments between 0.2 °C and 1.4 °C, with average increments of 0.3 °C for the spring–summer period. These increments were located above the 75th percentile.

For 2011 (Figure 6b), there were seven months with positive anomalies oscillating between 3% and 9% of the total range, implicating temperature increments between 0.3 °C and 2.0 °C. These increments were recorded during two months of the autumn–winter period (January and March) and five months of the spring–summer period, with average increments of 0.9 °C. These increments were located above the 75th percentile.

For 2013 (Figure 6c), February registered the highest temperature anomaly, 1.5 °C above the average, followed by October at 4% above the average and June and July at less than 3%. The other months did not register positive temperature anomalies.

Only March and April registered temperatures below the long-term average during 2019 (Figure 6d). The other months had positive anomalies from 1 to 6% above the media, with an average temperature increment of 0.7 °C for this year. During the spring–summer season, the temperature was 1.6 °C higher than the mean of the long-term period, corresponding to increments above the 75th percentile of the total period.

Figure 7 shows the monthly precipitation behavior during four years with Disaster Declarations by drought: 2005, 2011, 2013, and 2019 in relation to the long-term climatology (1980–2022). During 2005 (Figure 7a), the autumn–winter season (November, December, and January) and the spring–summer season (April and September) recorded negative anomalies. These anomalies oscillate between −10 and −80%, representing −5.7 and −54.6 mm of accumulative monthly rain. Nevertheless, October and February registered higher precipitations, up to 110% above the normal, representing 127.8 mm more than the average.

In 2011 (Figure 7b), negative anomalies were recorded in eight months, ranging from −20 to −70%, representing −6 to −100 mm of less accumulative rain. Positive anomalies were recorded in January, June, July, and November, ranking from +20 to +45% of the normal, representing up to 69 mm more accumulative rain during July, with the highest precipitation records. Nevertheless, the negative anomalies were below the 75th percentile in more than 50% of the year. The average reduced precipitation recorded for the spring–summer season was −126.2 mm, while the autumn–winter season registered −58.6 mm less than the normal. The total reduction in precipitation during 2011 was 18% lower than the average of the long-term period.

In 2013 (Figure 7c), eight months had positive anomalies, with precipitations from 10 to 120% higher than average, and November being the month with the highest precipitations. These positive anomalies represent up to 254 mm more accumulative rain for the spring–summer season and 104 mm for the autumn–winter season. The negative anomalies were recorded only during January, February, and April, with average reductions of −10 mm, representing between −5 and −25% less rain than the long-term average.

In 2019 (Figure 7d), 11 months registered negative anomalies, and only October registered positive ones. These values oscillated between −2 and −75%, with precipitations reduced between −1 mm and −102 mm. During this year, six months had accumulative rainfall below the 25th percentile. Therefore, 2019 registered a total reduction of 30% below the long-term average.

3.3. Corn Grain Production

Figure 8a,b shows each municipality’s average rainfed corn grain sown during the study period (2003–2022). The spring–summer season has the highest proportion of land, with 68% of the total surface of the year cultivated during this season. Below latitude 20°, the southern region has five municipalities cultivating more than 10,000 ha, while the northern has only one municipality during the spring–summer season. During the autumn–winter season, only one municipality cultivated more than 10,000 ha in the region. Figure 8c,d shows the volume obtained in the study region. Below parallel 21°, three municipalities produced more than 20,000 tons; above the parallel 22°, only three sites harvested more than 10,000 tons but less than 15,000 tons.

3.4. Drought Levels and Variations in Corn Grain Productivity

There were negative significant correlations between drought levels and yield values in 11 of 143 municipalities for the autumn–winter season (p < 0.01). In comparison, there were ten negative significant correlations of 231 municipalities between drought levels and yield for the spring–summer season (Figure 9a). Six affected municipalities were above the 22.5° parallel during spring–summer and four below the parallel 20°. During the autumn–winter season, only one site with negative correlations was below the parallel 21°. There were significant negative correlations between drought levels and volume values in four municipalities during the spring–summer season and 10 for the autumn–winter season (Figure 9b). During the spring–summer season, two sites were below parallel 19°; during the autumn–winter season, three sites with negative significant correlations were located below parallel 21°.

Correlations between corn grain yield variations and drought intensity registered nine statistically significant years for the spring–summer and eight for the autumn–winter season. Four years were statistically significant for the corn grain volume variation during the spring–summer season and five years for the autumn–winter season (

Table 3. Correlations between corn grain production variations and drought levels during 2003–2022.

Year	Spring-Summer Season				Autumn-Winter Season
	Yield		Volume		Yield		Volume
	r	p	r	p	r	p	r	p
2004	0.11	0.08	0.13	0.04 *	−0.41	0.00 **	−0.30	0.00 *
2005	0.22	0.00 **	−0.13	0.03 *	0.19	0.03	0.11	0.07
2006	−0.36	0.00 **	−0.16	0.01 *	0.19	0.02 *	−0.09	0.17
2007	−0.24	0.00 **	−0.07	0.30	−0.24	0.00 **	−0.09	0.13
2008	0.14	0.03	0.02	0.81	0.27	0.00 **	0.03	0.67
2009	0.15	0.03	0.12	0.05	0.09	0.32	0.13	0.03 *
2010	−0.28	0.00 **	−0.42	0.00 *	−0.02	0.81	−0.09	0.15
2011	−0.15	0.02 *	−0.24	0.00 *	−0.24	0.01 **	−0.07	0.26
2012	0.03	0.60	−0.13	0.05	−0.10	0.27	−0.07	0.29
2013	0.04	0.50	0.04	0.50	0.08	0.33	0.04	0.49
2014	0.15	0.02 *	0.13	0.04 *	−0.04	0.66	−0.04	0.49
2015	−0.02	0.75	−0.06	0.34	−0.21	0.01 **	−0.06	0.37
2016	0.23	0.00 **	0.12	0.06	0.03	0.67	0.01	0.84
2017	−0.02	0.76	−0.01	0.89	−0.19	0.02 *	−0.17	0.01 *
2018	−0.07	0.28	−0.16	0.01 *	−0.07	0.40	−0.14	0.02 *
2019	−0.24	0.00 **	−0.04	0.52	−0.01	0.88	0.05	0.45
2020	0.06	0.34	0.10	0.13	0.00	0.99	−0.13	0.04 *
2021	0.09	0.16	0.11	0.08	−0.18	0.03 *	−0.14	0.02 *
2022	0.29	0.00 **	−0.11	0.07	0.24	0.00 **	0.26	0.00 **

* Statistically significant (p < 0.05), ** statistically significant (p < 0.01).

). For the spring–summer season, in two years with positive correlations (2005 and 2014), the drought levels were below the total median of the 2003–2022 period. However, in 2016, 2020, 2021, and 2022, the yields experienced higher increments, independent of the drought levels. For the autumn–winter season, the years with negative correlations in corn grain yield corresponded with increments in the drought levels above the long-term average and vice versa. Similarly, the years with drought levels above the average corresponded with significantly different decrements in the production volume and vice versa (Figure 10).

3.5. Corn Grain Productivity Versus Drought Intensity

Figure 10 shows the performance of yield and volume production during the spring–summer and autumn–winter seasons versus the total drought intensity levels. In all cases, yield and volume decrease when drought intensity increases. The higher determination coefficient was obtained with linear regression, and the highest correlation was observed for the yield during the spring–summer season (Figure 11a). With the highest sum of drought intensities (14.33) during the spring–summer season, the yield can decrease to −1.77 tons∙ha⁻¹. This total drought intensity level represents D2 and D3 levels recorded during at least 50% of the period of six months. During the autumn–winter season (Figure 11b), the decrements are slightly lower, with up to −1.38 tons∙ha⁻¹ for average drought levels of D2. The volume production indicates that the autumn–winter season is more affected by drought intensity (Figure 11d), with 22% higher decrements during autumn–winter than during spring–summer (Figure 11c) and 11% lower intensity levels during the autumn–winter season.

3.6. Damaged Corn Surfaces and Drought Intensity

Figure 12 shows the damaged surfaces of the corn grain and the land uses from 2003 to 2022. During the spring–summer season, the years with Disaster Declarations by drought reported 86,449.88 ha lost; during the autumn–winter season, only 18,412.87 ha were lost. This data corresponds to 4.70 times more damaged surfaces in spring–summer than in autumn–winter. The two municipalities with the highest damaged surfaces during the spring–summer season (more than 10,000 ha) were Chicontepec, Veracruz, and Tula, Tamaulipas, located in warn-subhumid and semiwarm dry climates, respectively, both above the parallel 21°. During the autumn–winter season, only one locality reported more than 5000 ha, Tantoyuca, Veracruz, in a warm-subhumid climate. Four out of the 231 municipalities sowing corn grain were statistically significant for crop grain surface damages during the spring–summer season (p < 0.05), while five out of the 143 municipalities were significant for the autumn–winter season. These locations corresponded with the more extensive land surface used for agriculture. There is no statistical significance for surface loss and Disaster Declarations by drought during 2005, 2013, 2018, and 2019 in the autumn–winter season. Still, statistical significance exists between Disaster Declarations by drought for 2011 in the spring–summer season (r = 0.425, p < 0.001).

4. Discussion

4.1. Drought Levels in the Study Region

Since the Drought Monitor started recording drought data, drought has been registered in 91% of the municipalities during the spring–summer season and in 59% of the municipalities during the autumn–winter season. These results indicate that the study region is prone to drought, with the spring–summer season being the most sensitive to drought. The northern side of the area (above parallel 25°) has climates classified as warm-semidry and warm-dry [28]. Therefore, high average drought levels (D2, D3, and D4) should be considered as usual. Unfortunately, no functional meteorological stations were found at these latitudes to complement the analysis with the Drought Monitor. This situation emphasizes the need to consider the Drought Monitor as the only available tool to analyze the drought phenomena in the Mexican sites without available meteorological data.

In the south (below parallel 18°), three municipalities stand out for their high levels of average drought (D2), which should not be expected in a warm-humid climate, with average yearly precipitations between 2500 and 3500 mm [20]. They also do not have any available meteorological data [26]. Therefore, it is impossible to analyze these sites’ climatic behavior. In this region, there has been a continued productive reconversion of the original tropical forests of the site to the current grasslands dedicated to cattle [20]. Therefore, these authors consider that the increment of the grasslands and the consequent decrease in the tropical forests is causing a climate change, transforming the place from a humid climate to a subhumid climate with lower precipitation levels, as has been observed in other sites of the region that have experienced similar transformations [30].

4.2. Disaster Declarations by Drought and Their Relationship with Corn Grain Production

Since the beginning of available national statistics at the municipal level, 2005, 2011, and 2019 were the years with the highest drought periods during Mexico’s spring–summer and autumn–winter seasons [4]. At the same time, there were Drought Declarations in 2005, 2011, 2013, 2018, and 2019 in the study region, with 138 during the spring–summer and 167 during the autumn–winter season [25]. In this regard, the Disaster Declarations agree with the higher drought levels (D3 and D4) recorded for 2011, 2013, and 2019 for the municipalities registering corn grain losses due to drought during both seasons (spring–summer and autumn–winter). For 2005, the Mexican Drought Monitor reported lower drought levels (the D3 level was the highest). However, this year, 100% of the municipalities registered D0 or D1 levels below parallel 22.5° during January and April, 99% during February and May, and more than 57% during March and June [4]. Therefore, it can be considered that these three months affected by abnormal and moderate drought were catastrophic for corn grain production, increasing the damaged surface by 19% compared with the previous 2004 [11]. Complementing this information, the anomalies obtained with the climatological stations indicate that 2005 recorded higher temperatures than the average long-term data (Figure 6). Thus, these higher temperatures might have contributed to the damage caused by the droughts recorded in the 2005 national disaster records, especially during the autumn–winter season, when January registered temperatures above 75% of the long-term average, as seen in Figure 6. Higher temperatures and lower precipitations during December, January, April, and September (Figure 7) indicate that precipitation decrements affected both seasons, as shown in the Drought Monitor.

Unfortunately, the crops affected since 2011 were not recorded in the national agronomical statistics used in this research. The Mexican government replaced the direct economic support for agricultural damages with independently hired catastrophic insurance [25]. Under this new system, the insurance is only applied if there is 70% or more crop surface damage; otherwise, damaged croplands are not registered. This situation implies a subregister of data reported in the Mexican Agrifood Statistical System. For example, for 2019, the official records only indicate that corn was one of the crops affected by drought that received insurance payments. Still, they do not specify the names of the municipalities nor the surface receiving these insurance payments [31].

Table 4. Documented drought and corn production were affected from 2003 to 2022 in the study site.

Period	Site and Situation	Study Type and Reference
2019	Atypical drought caused crop damage in 15 municipalities in Veracruz. A national increment in corn importation was documented due to drought.	State Agrifood Report [31]. Mexican newspaper [32].
2018–2020	The municipality of El Mante, Tamaulipas, reported that 24% of its corn surface was loose due to a rain deficit.	Direct information from local producers [33].
2014	The ten most productive municipalities in Veracruz reported decrements in corn yield due to drought and a deficit in government economic support.	Data was obtained from the National System of Agrifood Information and the state of Veracruz [34].
2011–2012	The municipality of Tatahuicapan de Juárez, Veracruz, reported a total loss of corn production.	Data was obtained from the National System of Agrifood Information and the state of Veracruz [35].
2011	San Fernando, Jaumave, Abasolo, González, Díaz Ordaz, Victoria, and El Mante, Tamaulipas, reported damaged corn lands due to drought during the spring–summer season.	Information was obtained from a Mexican newspaper [36].
2005	Drought affected the northern side of Mexico and Veracruz, causing corn losses and affecting 950,000 ha of different crops.	Information was obtained from a Mexican newspaper [37].

shows a brief resume of data from different public sources regarding drought and its effects on corn production on the study site. As seen in the table, local data is scarce and incomplete. However, 2011 and 2019 are documented as years with total corn surface lost due to drought. This information confirms our study findings, which indicate that these municipalities reported corn grain surface damages during the Disaster Declarations due to drought.

4.3. Drought Levels and Corn Grain Yield and Volume

The number of significant correlations between yield and volume decrements and drought levels indicates that the spring–summer season has the most significant coefficient determinations (Figure 10). These results were obtained because more municipalities are sowing corn grain during this season; therefore, there were more substantial numbers of data (239 municipalities sown corn grain during the autumn-summer and only 151 during the autumn–winter season). For example, the droughts recorded during 2011 affected the spring–summer season more, with an average decrement of −695 kg ha⁻¹, compared to the autumn–winter season, which registered a decrement of −8.57 kg ha⁻¹. Low yield associated with drought conditions under rainfed corn has been previously documented for the eastern side of Mexico compared to irrigated corn production. However, there are no specific estimations of these decrements [38].

In Mexico, native corn varieties have been selected to be cultivated under the country’s different agroclimatic conditions, especially considering drought tolerance and fast-growing and better performance. These studies have found that drought mainly affects chlorophyll contents, independently of the genetic origin, affecting the root and biomass development [39]. This situation implies that no matter how many new corn varieties are developed to tolerate drought, this plant will always be highly vulnerable. One study reported a drought-tolerant variety of white corn grain produced in the northern side of the study region. The authors indicated that this corn obtained an average harvest period of 100–110 days during the spring–summer season and 120–125 days during the autumn–winter season. Under experimental conditions and 350 mm of precipitation during the plant cycle, this variety reached a maximum yield of 3.78 tons∙ha⁻¹ [40], above the average of the study site production of 2.44 tons∙ha⁻¹. Nevertheless, the experimental data indicate that yield can decrease by 1.75 tons∙ha⁻¹ under rainfed conditions compared with irrigated conditions in the same area.

Local research in the northern side of the region, under a warm subhumid climate, documented that the average rainfed corn grain yield is between 1.42 and 2.78 tons∙ha⁻¹ [33]. This research also indicates that the local producers use certified and native corn grains adapted to their natural climatological conditions. However, these producers also manifest that drought is the leading cause of their low productivity and crop losses. Conversely, in the southern part of the study region, there are municipalities with averages between 3.86 and 4.01 tons∙ha⁻¹ under a warm-humid climate [11]. These results indicate that precipitation is a crucial factor in incrementing yield. More rain and yield improvement have been demonstrated in specific studies where hydric stress played an important part when corn was flowering, decrementing its yield by 55%, even with a highly productive corn variety, when comparing a tropical humid forest versus a dry forest climate [41].

In a broad sense, national studies suggest that the minimal precipitation estimated for the Mexican corn was established at 480 mm during the plant’s life cycle. Lower precipitations are not recommended under rainfed conditions [12]. Therefore, the results found in this study for the rainfed corn grain sown at the study site indicate that these national studies are more congruent with the corn varieties used in the study region. Studies like the one mentioned previously in the northern side of the study region and others in similar climate types [33,41] reporting yields between 3.4 and 3.7 tons∙ha⁻¹ under drought conditions may need further review. The local statistics in this study indicate that producers may not apply the same care and techniques as the scientific reports document in their experiments. Additional studies document regional differences in production yields even with the same white corn variety. The regions with better yields have standard larger-scale production, improved germplasm, and intensive production systems. In comparison, the areas with lower yields have common environmental and substantial technological limitations [38].

Therefore, the study region may need more training in corn grain production and expert technical directions because national statistics indicate that 57% of these producers only finished elementary school (six years), and 14.8% do not have any education [42]. This situation makes it more difficult for corn cultivators to apply new technical developments and follow the instructions to use the new varieties developed by governmental research institutions for specific agroclimatic conditions [43].

The national statistics show that the average corn grain yield is 3.9 tons∙ha⁻¹. However, this data includes rainfed and irrigated corn production [13]. For example, in the state with the highest corn grain production (Sinaloa on the northern and western sides of Mexico), irrigated corn grain has an average yield of 10.9 tons∙ha⁻¹. In this same state, with high agricultural potential for corn, the rainfed corn averages 2.9 tons∙ha⁻¹ [11], only slightly higher than the study region’s average.

The other variable, corn grain volume (tons), is more related to the quantity of cultivated land during the period. Therefore, the municipalities with higher surfaces of corn sown during the driest seasons were the most affected and significantly correlated with drought levels reported by the Drought Monitor (Figure 8 and Figure 9). Municipalities like Las Choapas, on the southern side of the region (with an average seasonal surface sown of 13,300 ha), registered −223.7 and −578.5 tons in the spring–summer season of 2011 and 2019 compared with their previous years. Papantla, in the region’s center, with an average surface of 13,576 ha, recorded −3652 and −5966 tons in the same season of 2011 and 2019 [11].

Unfortunately, no meteorological stations were available for the Las Choapas and Papantla municipalities during the study period. Besides, a nearby Las Choapas site (Coatzacoalcos) did not contain rain precipitation data during 2011 but reported less than 30% of the normal rainfall from February to May 2019. Papantla’s nearest meteorological station registered only 41% of the average rain precipitation from March to June and 80% of the normal from October to December 2011. During 2019, this station recorded 42% of the normal rainfall from July to September, affecting mainly the spring–summer season [26]. Therefore, we consider that, even with these scarce meteorological data, it is observable that corn grain was affected by low precipitations during 2011 and 2019. The year 2019 represented the most significant volume decrements because 2018 also had high drought levels in both municipalities (D0 from May to June, D1 and D2 for July, and D3 for August and September for Las Choapas, while D0 from June to August and D1 in September for Papantla). Thus, implementing more meteorological stations in the study site to correlate corn productivity and climate data is advisable to obtain better estimations.

Nevertheless, these results correlate with the higher volume decrements estimated with the linear regression equation obtained in this study. However, this equation underestimates the volume reduction for these two cases and obtains a low correlation with the drought levels registered at the sites. The lower correlation and coefficient of determination obtained with the linear models are congruent with sown fluctuations between seasonal periods. In this case, on average, only 2005 and 2019 recorded significantly decreased production volume due to drought. However, 2011 was not negatively affected in some municipalities of the study site compared with the previous year during the spring–summer season because 2010 registered one category-three hurricane that impacted the study region’s center. However, in the significant municipalities, 2011 recorded 387.59 tons less than in 2010 during the spring–summer season and −269 tons during the autumn–winter season.

These peculiarities in corn production decrements and drought intensities indicate that other phenomena besides drought may affect corn grain production in the study region. This region is also affected by snow, hail, strong winds, excessive rains, floods, and tropical cyclones [25], which makes it more challenging to estimate drought and corn production correlations. Therefore, only the local meteorological stations can provide more reliable data to analyze the corn grain performance related to drought levels.

5. Conclusions

The North American Drought Monitor and its Mexican part, the Mexican Drought Monitor, can be a helpful tool to predict corn grain productivity behavior in specific sites without local meteorological data and when other phenomena, like tropical cyclones, floods, or excessive rains, were not registered. Under these situations, drought intensities with levels D3 (extreme drought) and D4 (exceptional drought) during the spring–summer season and autumn–winter season may cause a total loss of corn grain production, mainly in warm-subhumid and temperate-subhumid climates. During the spring–summer season, crop surface damage can be 4.7 times higher than during the autumn–winter season due to the more considerable land extension cultivated during this season. During the spring–summer season, corn grain yield can decrease by 1.77 tons∙ha⁻¹ and 1.38 tons∙ha⁻¹ during the autumn–winter season with D2 and D3 drought levels. The spring–summer season has a higher predictability than the autumn–winter season due to the higher number of municipalities and land surfaces cultivating this crop during this period, which provides a more considerable amount of data to correlate drought and corn grain production.

We suggest more detailed in situ studies of drought levels and corn grain production, considering corn varieties and cultivation techniques, to improve the predictability of the drought effects on corn grain production under rainfed conditions.