Assessment of Census and Remote Sensing Data to Monitor Irrigated

Agriculture in Mexico

Jean-Francois Mas

1 a

and Azucena P

erez-Vega

2 b

Centro de Investigaciones en Geograf

ıa Ambiental, Universidad Nacional Aut

onoma de M

exico, Morelia, Mexico

Departamento de Geom

atica e Hidra

ulica, Universidad de Guanajuato, Guanajuato, Mexico

Keywords:

Irrigated Agriculture, Evapotranspiration, Water, Remote Sensing.

Abstract:

Irrigated agriculture faces imminent threats, such as escalating water scarcity and climate change impacts. Wa-

ter scarcity is particularly crucial in countries such as Mexico, where approximately 41% of the land comprises

arid and semi-arid zones. The study assesses the quality and consistency of monitoring irrigated agriculture

in a municipality of the State of Guanajuato in central Mexico using agricultural census information and ad-

vanced remote sensing data from Landsat 8, MODIS, and ECOSTRESS. Preliminary analyses showcase the

dominance of wheat and barley crops in P

enjamo, with MODIS time series effectively capturing crop growth

dynamics. The study highlights the potential of remote sensing in estimating irrigated crop dynamics propor-

tions and the associated water consumption at different scales.

1 INTRODUCTION

Irrigated agriculture is crucial to food and economic

security in many countries worldwide. However, this

practice has faced many signiﬁcant challenges in re-

cent decades that threaten its long-term sustainability.

Two of the most pressing problems facing this sector

are increasing water scarcity and the effects of climate

change. The combination of these factors presents a

worrying picture that demands a deeper and more de-

tailed understanding of the dynamics of irrigated agri-

culture.

The National Commission of Arid Zones

(CONAZA) reports that Mexico has around 41% arid

and semi-arid zones. Water scarcity has emerged

as a fundamental obstacle to irrigated agriculture in

this context. As water demand increases for both

agricultural and urban use, available water resources

are reaching critical levels. This imbalance between

water supply and demand raises crucial questions

about the long-term viability of irrigated agriculture,

which has historically relied heavily on water sources

now threatened by overexploitation and climate

variability (Madramootoo and Fyles, 2010). Climate

change adds complexity to this problem, affecting

precipitation patterns and increasing the frequency

https://orcid.org/0000-0002-6138-9879

https://orcid.org/0000-0002-9683-4207

and intensity of extreme weather events (Hanjra and

Qureshi, 2010).

In Mexico, the two agricultural cycles, au-

tumn/winter and spring/summer, play a crucial role

in the country’s food production. During the au-

tumn/winter, irrigated agriculture is relevant since

rains are scarce and irrigation systems primarily de-

pend on guaranteeing the necessary water supply for

crops. In contrast, the spring/summer cycle coincides

with the rainy season, which reduces the need for irri-

gation in certain regions (see Figure 1). Understand-

ing the dynamics of irrigated agriculture and asso-

ciated water consumption is crucial for efﬁcient wa-

ter resource management and climate change adapta-

tion strategies. Addressing these challenges requires

a multidisciplinary approach that integrates scientiﬁc,

technological and policy knowledge to ensure the sus-

tainability of irrigated agriculture in a changing en-

vironment. In this work, we will limit ourselves to

evaluating some ofﬁcial census and remote sensing

inputs to monitor irrigated agriculture in a municipal-

ity in central Mexico. The aim is to evaluate the qual-

ity and consistency of the information obtained from

these various sources.

Mas, J. and Pérez-Vega, A.

Assessment of Census and Remote Sensing Data to Monitor Irrigated Agriculture in Mexico.

DOI: 10.5220/0012701500003696

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 10th International Conference on Geographical Information Systems Theory, Applications and Management (GISTAM 2024), pages 181-186

ISBN: 978-989-758-694-1; ISSN: 2184-500X

181

Figure 1: Ombro-thermal diagram of P

enjamo,

Guanajuato. (Data obtained from CONAGUA

https://smn.conagua.gob.mx).

2 STUDY AREA

The State of Guanajuato is located in central Mexico

(Figure 2). The historical evolution of Guanajuato’s

agricultural land can be traced back to pre-Hispanic

times, but a signiﬁcant increase in croplands occurred

during the colonial period (P

erez-Vega, 2011). Nowa-

days, Guanajuato is among the Mexican states with

the highest percentage of transformed covers (around

60%), including irrigated agriculture (21.5%), rain-

fed agriculture (25.7%) and pasture (9.7%) (Palacios

et al., 2000).

Figure 2: State of Guanajuato in Mexico (red polygon). The

study area (P

enjamo) corresponds to the white square.

Regarding water resources management, ground-

water supplies 70% of human needs; however, all

aquifers in the state register levels of overexploitation.

Most of the water is destined for agricultural activity.

Water availability in the state is below scarcity, with

845 m

/inhabitant/year, while the national average is

3,705 m

/inhabitant/year (IEE, 2008).

3 MATERIAL

3.1 Agricultural Census and Climate

Data

Data from the Agri-Food and Fisheries Information

Service (SIAP) provide valuable information on agri-

cultural activities in Mexico (http://infosiap.siap.gob.

mx/gobmx/datosAbiertos.php). These open-access

data are presented in tables and offer details on the

cultivated, harvested and damaged areas by crop cy-

cle and production mode (rain-fed / irrigation) and

are grouped according to spatial units such as states,

municipalities and irrigation districts (from 2003).

The information provided by the SIAP offers a de-

tailed vision of agriculture in Mexico during the

last decades, facilitating agricultural sector decision-

making (P

erez-Vega and Mas, 2023).

We also used data from the National Water Com-

mission (CONAGUA) meteorological stations.

3.2 Cartography

We used the land use and vegetation map, scale

1:250,000 from INEGI, the ofﬁcial cartography and

statistics agency of Mexico, to delimit the irrigation

areas and a map of the municipalities (INEGI) to spa-

tialize the agricultural census data.

3.3 Remote Sensing Data

We used Landsat 8 multispectral imagery. These im-

ages have nine spectral bands (spatial resolution of

30 meters), including a panchromatic band with 15 m

resolution. The temporal resolution of 16 days and the

existence of a historical collection with images with

this spatial resolution since the 1980s facilitates the

monitoring of changes in land cover and the detection

of environmental phenomena over time.

The MOD13A1 product, version 6.1, derived from

data collected by the MODIS (Moderate Resolution

Imaging Spectroradiometer) sensor, focuses on global

vegetation monitoring. It offers a spatial resolution of

500 meters, allowing evaluation of large-scale vegeta-

tion patterns with a frequency of 16 days since 2000.

Among its notable features is the ability to calculate

the Normalized Difference Vegetation Index (NDVI),

a crucial metric for assessing the health and density

of vegetation in different regions of the planet.

Another derivative product of the MODIS sensor

is the MOD16A2 (version 6), explicitly designed to

estimate water balance on a global scale. This prod-

uct provides detailed information on evapotranspira-

tion (ET) and water availability at the Earth’s surface.

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

182

With a spatial resolution of 500 meters and for periods

of 8 days, the MOD16A2 allows a detailed analysis of

water consumption patterns at a regional level, thus

contributing to the sustainable management of water

resources.

The ECOSTRESS sensor (Ecosystem Spaceborne

Thermal Radiometer Experiment on Space Station)

has been developed to measure the temperature

of the Earth’s surface with high precision. The

ECO3ETPTJPL product provides detailed data on In-

stantaneous Latent Heat Flux (W/m

), which is di-

rectly related to evapotranspiration, at a spatial resolu-

tion of approximately 70 meters. This high resolution

allows the study of speciﬁc phenomena at the local

and regional level, providing valuable information to

understand hydrological processes and the response

of ecosystems to changes in water availability.

We used the Application for Extracting and Ex-

ploring Analysis Ready Samples (AppEEARS, avail-

able at https://appeears.earthdatacloud.nasa.gov/) tool

to obtain ECOSTRESS evapotranspiration values.

Climatic data were obtained and processed with the

Climex program (

Angel Marqu

es-Mateu et al., 2023).

We performed all other statistical analysis, mapping

and graphing operations with the R program (R Core

Team, 2021).

4 METHODS

4.1 Preliminary Analysis of Census

Data

The SIAP data were analyzed to determine which mu-

nicipalities have important irrigation areas (more than

10,000 ha) dominated by a single crop (more than

80%, during 2003-2023) and whether a change in

dominant crop occurs. In this preliminary study, we

sought a more straightforward case to avoid dealing

with an area with different crops with different spec-

tral responses and planting and harvesting schedules.

4.2 Analysis of Agricultural Dynamics

and Estimation of the Sown Area

We prepared time series with MODIS (ET and NDVI,

2002-2023) and ECOSTRESS (Latent Heat Flux,

2018-2020) data. We selected 30 random points in ir-

rigation areas and drew graphs showing the temporal

variations of these variables, allowing us to observe

the areas sown or not in the two cycles.

In the next step, we evaluated to what extent re-

mote sensing data at different spatial resolutions allow

evaluation of the sown area during the winter/autumn

cycle. We computed the correlation coefﬁcient be-

tween the sown area reported by the SIAP and in-

dices obtained from the images as the sum of the

NDVI and ET values in the irrigation areas during

the autumn/winter cycle. An unmixing exercise of

the MODIS images was also carried out to estimate

the sown area obtained from the Landsat image for

the same date.

To estimate the evapotranspiration of croplands,

we used the Blaney-Criddle method (Blaney and

Criddle, 1950) because it is accurate enough and only

requires temperature data (Equation 1).

= K

∑

p(0.457T

mean

+ 8.128) (1)

Where: Et

is the crop evapotranspiration (mm),

mean

is the mean daily temperature, p is the mean

daily percentage of annual daytime hours, and K

the crop coefﬁcient.

We obtained the temperature from data from cli-

matological stations for 1981-2010. We chose sta-

tions that best represented the climatic conditions of

irrigated areas. We used crop coefﬁcients from local

and international sources (Allen et al., 1998;

Angeles

Hern

andez et al., 2017; INIFAP, 2001). We compared

these ET values with those obtained from the analysis

of MODIS and ECOSTRESS images.

5 RESULTS

One municipality that met the search criteria was

enjamo in the State of Guanajuato. The dominant ir-

rigated crop (more than 80% during 2002-2014) was

wheat, and from 2016, wheat and barley. The to-

tal irrigation area sown during 2003-2022 varied be-

tween 12,000 and 34,500 ha (autumn-winter cycle)

and 18,000 and 37,000 ha (spring-summer cycle).

As shown in Figures 3 and 4, MODIS time series

(NDVI and ET) allow crop growth to be clearly ob-

served in both cycles.

The ECOSTRESS data obtained for the same

points were very noisy. We observed very high la-

tent ﬂux values, which the uncertainty layer did not

allow to eliminate (Figure 5).

The NDVI was calculated based on a Landsat 8

image from March 2, 2020, when the contrast be-

tween cultivated and non-cultivated areas appeared

strongest. We determined visually that the value of

0.3 separated the areas with crops. The binary crop

map obtained by thresholding indicates a cultivated

area of 13,650 ha, while the SIAP reports a very close

value (13,905 ha). The Landsat binary crop map was

overlaid with the MODIS NDVI image of the same

Assessment of Census and Remote Sensing Data to Monitor Irrigated Agriculture in Mexico

183

Figure 3: Variation of evapotranspiration (obtained from

MODIS) over a year on some sampling points of the irriga-

tion area of P

anjamo, Guanajuato. Each color represents a

year on the period 2003-2022, time is shown as julian days.

Figure 4: Variation of mean NDVI (obtained from MODIS)

over a year on some sampling points of the irrigation area

of P

anjamo, Guanajuato. Each color represents a year on

the period 2003-2022, time is shown as julian days.

Figure 5: Variation of latent heat ﬂux (obtained from

ECOSTRESS) over 2019-2022 on the same sampling points

of the irrigation area of P

anjamo, Guanajuato. Time is

shown as julian days from 1st Jan. 2019.

period (MOD13A1 NDVI 2020 081), thus allowing

the proportion of crops in each 500 m MODIS cell

to be calculated. The correlation coefﬁcient between

NDVI and crop proportion was 0.85, which suggests

that NDVI is a good indicator of crop proportion (Fig

6). We ﬁtted a linear model to explain the crop ratio

with the NDVI value. The model presented a good ﬁt

(Adjusted R-squared = 0.82).

Figure 6: Relationship between the proportion of irrigated

crop in coarse MODIS pixels and NDVI values.

We sought to estimate the relationship between the

NDVI and ET values from MODIS on the one hand

and the total area sown according to the SIAP during

2003-2022. To do this, we calculated the sum of each

year’s NDVI and ET values of the irrigation district

cells. We evaluated the relationship between these

values and the sown areas reported by the SIAP. Sim-

ilarly, the unmixing model reported in the previous

paragraph for each year (Figures 7, 8 and 9).

Figure 7: Relationship between the yearly swon irrigated

area (autumn/winter cycle) during 2003-2022 from SIAP

and the sum of MODIS ET values.

The best relationship was obtained using the evap-

otranspiration (R

=0.49). However, it is not a strong

relationship and does not allow us to estimate the

snow-irrigated area accurately. These limitations have

reported in the literature (Wu et al., 2022).

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

184

Figure 8: Relationship between the yearly swon irrigated

area (autumn/winter cycle) during 2003-2022 from SIAP

and the sum of MODIS NDVI values.

Figure 9: Relationship between the yearly swon irrigated

area (autumn/winter cycle) during 2003-2022 from SIAP

and the values obtained by unmixing MODIS NDVI values.

6 DISCUSSION AND

CONCLUSION

This study focuses on evaluating the quality and con-

sistency of information related to irrigated agriculture

in the State of Guanajuato, Mexico, with a particu-

lar emphasis on the municipality of P

enjamo, located

in the central part of Mexico. The methods include

analysing agricultural census data, cartography, and

remote sensing data from various sensors, such as

Landsat 8, MODIS, and ECOSTRESS. The research

speciﬁcally focused on P

enjamo due to its relevance

in irrigated agriculture and the challenges posed by

water scarcity in the region.

The results demonstrated the effectiveness of

MODIS time series in clearly depicting crop growth

in both cycles. However, ECOSTRESS data exhibited

high noise levels, impacting the clarity of latent ﬂux

values. In the future, we will also use Sentinel-2 data,

which presents high spatial and temporal resolution.

Because different modalities of data can complement

each other by combining their strengths and reducing

their limitations, Multimodal remote sensing (MRS)

methods are beneﬁcial for crop monitoring and are

gaining popularity (Karmakar et al., 2024).

The ﬁndings of this preliminary study provide

valuable insights into the dynamics of irrigated agri-

culture. The effectiveness of MODIS time series in

monitoring crop growth during different agricultural

cycles highlights the utility of coarse remote sensing

data in assessing irrigated agriculture over large ex-

tensions (e.g. the entire Mexican territory).

These ﬁndings contribute to the broader discus-

sion on applying multidisciplinary approaches, com-

bining agricultural census data, cartography, and re-

mote sensing, to address the sustainability of irrigated

agriculture in regions facing water scarcity and cli-

mate variability. The research underscores the impor-

tance of integrating different data sources and tech-

nologies to understand irrigated agriculture dynam-

ics comprehensively. As water scarcity intensiﬁes

globally, the methods and insights presented in this

study can inform policymakers in developing effec-

tive strategies for sustainable water resource manage-

ment and climate change adaptation in irrigated agri-

culture.

ACKNOWLEDGEMENTS

This study was carried out within the scope of

the PAPIIT-UNAM project (IN112823) entitled:

Azolvamiento y eutroﬁcaci

on en presas periurbanas

de zonas templadas de M

exico: contribuciones para

su evaluaci

on y prospecci

on and An

alisis de la Cali-

dad, Cantidad, Gesti

on y Pol

ıticas P

ublicas ante es-

cenarios de Cambio Clim

atico en la Presa Sol

ıs, Gua-

najuato, Direcci

on de Apoyo a la Investigaci

on y el

Posgrado, Universidad de Guanajuato).

REFERENCES

Allen, R., Pereira, L., Raes, D., and Smith, M. (1998).

Crop evapotranspiration – guidelines for computing

crop water requirements. FAO Irrigation and drainage

papers 56, FAO, Roma, Italy.

Blaney, H. and Criddle, W. (1950). Determining water re-

quirements in irrigated areas from climatological and

irrigation data. Soil Conservation Service Technical

Papers 96, US Department of Agriculture, Washing-

ton, D.C.

Hanjra, M. A. and Qureshi, M. E. (2010). Global water

crisis and future food security in an era of climate

change. Food Policy, 35(5):365–377.

IEE (2008). Informe ambiental del Estado de Guanajuato.

Technical report, Instituto de Ecolog

ıa del Estado.

Assessment of Census and Remote Sensing Data to Monitor Irrigated Agriculture in Mexico

185

INIFAP (2001). Requerimientos h

ıdricos de especies an-

uales y perennes en las zonas media y altiplano de san

luis potos

ı. Folleto t

ecnico 12, Instituto Nacional de

Investigaciones Forestales, Agr

ıcolas y Pecuarias.

Karmakar, P., Teng, S. W., Murshed, M., Pang, S., Li, Y.,

and Lin, H. (2024). Crop monitoring by multimodal

remote sensing: A review. Remote Sensing Applica-

tions: Society and Environment, 33:101093.

Madramootoo, C. A. and Fyles, H. (2010). Irrigation in the

context of today’s global food crisis. Irrigation and

Drainage, 59(1):40–52.

erez-Vega, A. and Mas, J.-F. (2023). An

alisis espacio-

temporal de la agricultura en M

exico: datos relevantes

del SIAP 1980-2020. In XXV Congreso SELPER, La

Paz BCS, 26-27/10/2023.

R Core Team (2021). R: A Language and Environment for

Statistical Computing. R Foundation for Statistical

Computing, Vienna, Austria.

Wu, B., Zhang, M., Zeng, H., Tian, F., Potgieter, A. B., Qin,

X., Yan, N., Chang, S., Zhao, Y., Dong, Q., Boken, V.,

Plotnikov, D., Guo, H., Wu, F., Zhao, H., Deronde, B.,

Tits, L., and Loupian, E. (2022). Challenges and op-

portunities in remote sensing-based crop monitoring:

a review. National Science Review, 10(4):nwac290.

Angel Marqu

es-Mateu, P

erez-Vega, A., and Molada-Tebar,

A. (2023). Climex. https://pypi.org/project/climex/.

Angeles Hern

andez, J., Trejo Segura, V., Unland Weiss, H.,

Olvera Salgado, D., and Carrillo Castillo, M. (2017).

La eﬁciencia de aplicaci

on y productividad del agua

en los cultivos de trigo y cebada en el distrito de riego

011 “alto r

ıo lerma”, guanajuato. In III Congreso Na-

cional de Riego y Drenaje COMEII 2017.

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

186