Modeling evapotranspiration in diverse climatic zones of Pakistan using Surface Energy Balance Algorithm for Land (SEBAL) through geospatial technologies

Abstract

Evapotranspiration (ET) is a key driver of the hydrological and energy cycles, influences directly the vegetation health, surface temperature regulation and the availability of water. Pakistan is a rapidly urbanized and climate vulnerable country, where, it is crucial to understand ET dynamics for sustainable land and water resource management. Landsat 7, 8 and 9 imageries and meteorological data from the Pakistan Meteorological Department were integrated using the Surface Energy Balance Algorithm for Land (SEBAL) in ERDAS Imagine to estimate actual evapotranspiration (ET). Key parameters, including Land Surface Temperature (LST), Normalized Difference Vegetation Index (NDVI), Albedo, and Net radiation, were derived and incorporated to assess spatiotemporal ET variations across Pakistan’s major climatic zones. The spatiotemporal analysis revealed a consistent rise in LST over the years 1989, 1999, 2009 and 2019, accompanied by a decline in NDVI after 2009, decreasing surface albedo, and fluctuating net radiation, indicating intensified urbanization, surface degradation, and altered energy dynamics across Pakistan’s climatic zones. Estimated ET values ranged from 4.67 to 7.33 mm/day, with a pronounced decline in the very high (6.94–7.33 mm/day), high (6.55–6.93 mm/day), and moderate (6.16–6.54 mm/day) ET zones, alongside a substantial expansion of the low (5.77–6.15 mm/day) and very low (4.67–5.76 mm/day) zones by 2019, highlighting extensive vegetation degradation, accelerated urban expansion, and increasing climatic stress across Pakistan’s diverse landscapes. The findings demonstrate the strong applicability and reliability of the SEBAL model for long-term environmental monitoring in data-scarce regions, highlighting its potential not only to guide sustainable urban development and strengthen climate resilience, but also to support integrated land–water management, inform adaptive policy frameworks, and advance scientific understanding of land–atmosphere interactions in the context of Pakistan’s evolving climatic and urban challenges.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-39756-1.

Introduction

Evapotranspiration (ET) encompasses both the evaporation of water from surface landscapes and the transpiration of moisture through vegetation and is a vital component of the Earth’s hydrological and energy cycles^1–3. It plays a crucial role in moderating land surface temperatures, supporting ecological balance and sustaining water availability^4,5. By absorbing latent heat during the vaporization of water, ET acts as a natural cooling mechanism, helping to offset the warming effects associated with rising global temperatures^6–8. However, in urban settings, this process is significantly disrupted by the expansion of impervious surfaces and the reduction of vegetative cover. These changes limit ET potential and contribute to the intensification of the Urban Heat Island (UHI) effect. As a result , these changes lead to elevated temperatures , higher energy demands , poor air quality and increased risks to public health^9,10. Pakistan’s diverse climatic landscape, from the arid deserts of the south to the temperate zones in the north, stimulates complexity to the understanding of ET dynamics^11,12. The country’s rapidly transforming landscapes, marked by expanding built-up, shrinkage of vegetation and land use pattern changes, indicates the pressure on natural ecosystems and processes of surface energy balance^13,14. These challenges are further intensified by strong seasonal contrasts, such as extreme pre-monsoon heat and variable monsoon humidity, that influence ET variability across both time and space^15,16. Traditional methods for estimating ET, while well-established, are often limited by their reliance on extensive ground-based data and localized field observations, making them less practical for large-scale or data-scarce regions. SEBAL provides high-resolution ET estimates approximately 85 % accuracy at the farm scale, while accuracy exceeds 95 % at the regional scale^17,18. Unlike global ET products such as MODIS Global Terrestrial Evapotranspiration (MOD16), Global Land Evaporation Amsterdam Model (GLEAM), the remote sensing technologies, particularly the Surface Energy Balance Algorithm for Land (SEBAL) approach provides good resolution and spatially explicit ET estimates and offers a robust and scalable alternative¹⁹. This enables more accurate, site specific assessment of water use, irrigation management and climate impacts across Pakistan’s diverse landscapes. SEBAL has been applied extensively across various geographical and climatic contexts from arid to semi-arid and even for humid regions not only for estimating ET but also for surface energy fluxes, monitoring agricultural water use, assessing drought impacts, managing irrigation efficiency and for analyzing urban heat dynamics^5,20. Developed by Bastiaanssen and further refined by Allen, SEBAL integrates satellite imagery with meteorological data to estimate ET with high spatial and temporal resolution^21–23. Given these advantages, SEBAL, offers a practical and efficient approach for monitoring ET over extended periods and diverse landscapes. The model’s ability to operate with minimal filed calibration and its integration of remote sensing makes it feasible to assess surface energy fluxes, vegetation dynamics and water balance²⁴ This makes SEBAL a valuable tool for understanding climate variability, supporting irrigation planning and formulating sustainable land and water management strategies in data scarce countries²⁵ like Pakistan. Also its adaptability and low data dependency make it particularly suitable for applications in Pakistan, where ground-based data are often limited^26,27. Building on this capability, this study aims to analyze the spatiotemporal variation in evapotranspiration (ET) and surface energy balance across the climatic zones of Pakistan from 1989 to 2019 using remote sensing and SEBAL modeling. It seeks to examine how changes in land use, vegetation covers and climatic conditions influence ET dynamics and surface temperature patterns. This study hypothesizes that increasing urbanization and land degradation have led to decline in ET and a ride in LST across climatic zones of Pakistan. Therefore, this study examines key surface energy balance components such as Land Surface Temperature (LST), Net Radiation (NR), Albedo and vegetation indices including Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI) in relation to evapotranspiration dynamics. These surface energy balance components play a vital role in SEBAL based ET estimation and help in assessing spatial variations in vegetation health and water use efficiency across Pakistan diverse climatic zones. Previous studies in Pakistan have focused on SEBAL-based evapotranspiration estimation in agricultural areas, with limited attention to its spatial and temporal variability across different climatic zones. This study fills that gap by modeling three decades of ET across Pakistan’s diverse climates using SEBAL and geospatial data, providing new insights for sustainable land, water and climate management. The novelty of this study lies in its comprehensive, nationwide SEBAL based assessment of evapotranspiration across five major climatic zones of Pakistan. By integrating multi decadal remote sensing data with climatic zone analysis, the study provides new insights into spatial- temporal ET dynamics in relation to urbanization and climate variability, supporting evidence based planning and sustainable resource management.

Materials and methods

Study area

Pakistan, with a land area of 880,940 km², is geographically diverse, ranging from the Indus plains to mountainous northern regions. Rapid urbanization, particularly in cities like Karachi shows an 80% population increase between 2000 and 2010²⁹, poses significant environmental challenges. With urban growth estimated at 3% annually, highest in South Asia, peri-urban expansion is reshaping the landscape²⁸. Climatically, Pakistan spans arid , semi-arid zones, humid and alpine zones with temperatures varying widely from glacial north to desert south³⁰. Due to this diversity, the study area is divided into five major climatic zones based on the classification scheme proposed by Qasim et al.³¹. Spatial representation of these climatic zones was produced by the authors using ArcGIS Basic Single Use License 10.5.1 (ESU 257390079) illustrated in Figure 1 and detailed in Table 1.

Fig. 1.

Study area map showing the climatic zones of Pakistan and its geographical location within South Asia and the world. The map was generated by the authors using ArcGIS Basic Single Use License 10.5.1 (ESU 257390079); https://www.esri.com). The classification scheme of climatic zones was adopted from Qasim et al.³¹.

Table 1.

Climatic zones of Pakistan.

Climatic zones	Description and features
Zone A	This zone is with cold climatic conditions and high elevated mountains. It is situated in the North of Pakistan located between 34 N o to 38 No and 70 E o to 78 E o in the Hindukush Himalayan ranges.
Zone B	The zone is located between 31 N o to 34 N o and 70 E o to 75 Eo with mild climatic conditions having sub-mountains.
Zone C	This zone is located between 27 N o to 32 N o and 64 E o to 70 Eo. It is a mountainous zone with high elevation and here the weather is cold in winters and hot in summers.
Zone D	The hottest and dry climatic zone of Pakistan having high mean temperatures located between 27 N o to 31 N o and 67 E o to 7 Eo.
Zone E	It is the biggest zone of Pakistan with coastal cities and the Arabian Sea. The coastal region is subdivided into small portion of Baluchistan province and arid Sindh province. It is located between 24 N o to 31 N o and 61 E o to 71 Eo.

Data acquisition

Cloud-free satellite imagery from Landsat 5 TM, Landsat 7 ETM+, and Landsat 8 OLI–TIRS (1989–2019) was obtained from the USGS Earth Explorer platform (https://earthexplorer.usgs.gov), covering Path/Row 157/40 to 145/34. The study period (1989,1999, 2009,2019) was chosen for its cloud-free Landsat TM, ETM+ and OLI data availability, ensuring consistent temporal analysis of ET estimation across Pakistan. All images were projected to UTM Zone 43N using WGS 1984. Preprocessing and analysis were performed using ERDAS Imagine and ArcGIS 10.5.1. All Landsat imageries were cloud-masked (<10% cover) using Fmask algorithm to remove clouds and cloud shadows ensuring reliable and consistent surface observations across the multi tile datasets. To address the Landsat 7 SLC-off data gaps (~22% missing pixels), focal analysis-based gap filling was applied in ERDAS Imagine 2014 using iterative spatial enhancement, effectively restoring image continuity and ensuring suitability for further analysis. The selected model SEBAL offers high reliability in data-scarce regions like Pakistan, requiring minimal ground data while effectively integrating satellite and meteorological inputs to estimate actual evapotranspiration across the country’s diverse climatic zones. Meteorological variables including air temperature, humidity, solar radiation and wind speed were acquired from the Pakistan Meteorological Department (PMD), Islamabad, to support SEBAL-based evapotranspiration estimation. Missing ground data were managed through temporal averaging and spatial interpolation using nearby station values to ensure data continuity and model reliability. These variables, then, were used as input parameters for the reference evapotranspiration calculation in the SEBAL model. These ground-based observations provided essential boundary conditions to calibrate satellite –derived surface energy fluxes, ensuring accurate estimation of actual ET across Pakistan’s climatic zones. Table 2 presents a comprehensive overview of the datasets utilized in this study.

Table 2.

Salient characteristics of satellite images deployed in the study.

Satellite	Sensor(s)	Path/Row	Year	Spatial resolution	Spectral bands	Temporal resolution
Landsat 5	Thematic Mapper (TM)	157/40 145/34	1989	30 m (visible, NIR, SWIR); 120 m (thermal; later 60 m)	7 bands	16 days
Landsat 7	Enhanced Thematic Mapper Plus (ETM+)	157/40 145/34	1999 2009	30 m (visible, NIR, SWIR); 15 m (panchromatic); 60 m (thermal)	8 bands	16 days
Landsat 8	Operational Land Imager (OLI) and Thermal Infrared Sensor (TIRS)	157/40 145/34	2019	30 m (visible, NIR, SWIR); 15 m (panchromatic); 100 m (thermal)	11 bands	16 days

Data analysis

A structured methodology was applied for data processing, including preprocessing, correction, and analysis. Landsat imagery (L5 TM, L7 ETM+, and L8 OLI-TIRS) from 1989 to 2019 was processed in ERDAS Imagine 2014, using the ACTOR (Atmospheric and Topographic Correction) module. This method was chosen for its reliability in minimizing atmospheric and topographic effects on Landsat imagery ^32,33. Radiometric calibration converted digital numbers to TOA reflectance³⁴, while LST was derived from thermal bands after emissivity adjustment³⁶. Cloud masking, alignment, and sub setting were applied³⁶, and final analyses were conducted using ArcMap 10.5.1 with Google Earth and Excel used for verification³⁷.

Evapotranspiration estimation using SEBAL

SEBAL is a remote sensing-based model used to estimate actual evapotranspiration (ETₐ) by solving the surface energy balance equation using satellite imagery and ground meteorological data. It computes the energy fluxes at the land surface through a structured series of steps.

Land Surface Temperature (LST)

LST, influencing the estimation of sensible and soil heat fluxes of SEBAL, was derived from the Thermal Infrared (TIR) band of Landsat data by using Planck’s equation³⁵.

1

Here, St stands for Land Surface Temperature, T is brightness temperature, λ is wavelength of the thermal band and ε represents the emissivity of land surface. The brightness temperature (T) was calculated using the following equation³⁸.

2

Here K2 and K1 are constants (band specific) representing difference radiance, after that Emissivity based on NDVI (Normalized Surface Vegetation Index) was calculated using Equation 3.³⁹, in Equation 3, PV ( Proportion to vegetation) indicates the vegetation percentage which was derived using NDVI following Equation 4.^28,39.

3

4

Surface albedo calculation

In SEBAL, surface albedo (α) was derived by converting satellite digital numbers to spectral radiance, then to reflectance, and finally summing weighted reflectance across bands. The top-of-atmosphere albedo is then corrected for atmospheric effects using Equation 5⁴⁰.

5

This value represents the surface’s shortwave reflectivity and is essential for net radiation estimation.

Vegetation indices calculation

In SEBAL, NDVI and EVI (Enhanced Vegetation Index) were used to assess vegetation cover and condition using following equations of 6⁴¹ and 7⁵ respectively:

6

EVI enhances sensitivity in high biomass areas and reduces atmospheric effects:

7

Where G is Gain factor, C1 and C2 are coefficient for red and blue p and L is Canopy background adjustment. These vegetative indices were used to compute vegetation fraction, surface emissivity and albedo which are critical SEBAL inputs for estimating the latent heat flux and, consequently, actual ET.

Net radiation (Rn)

Rn was calculated by Equation 8⁴² adding all incoming radiant fluxes Rs↓ (shortwave and longwave) with all outgoing radiant fluxes RL↑ (shortwave and longwave) while α represents surface albedo and outgoing and ε is surface emissivity.

8

Soil heat flux

Soil heat flux estimated as a fraction of Rn and represents the energy conducted into the It is calculated using surface temperature (Ts), albedo (α), and NDVI. Soil heat flux was calculated using Equation 9⁴³:

9

Where Tₛ is the surface temperature (°C), α is the surface albedo.

Sensible heat flux (H)

Sensible heat flux (H) was estimated using the Equation 10⁴³:

10

Where ρ (air density), Cp is specific heat of air, dT is the temperature difference between two surface heights, and rah is aerodynamic resistance. The values of dT and rah were derived using hot and cold anchor pixels to represent surface-atmosphere thermal gradients.

Latent heat flux

The latent heat flux (λET) in SEBAL was computed as the residual of the surface energy balance using Equation 11⁴³:

11

Where Rn is Net radiation (W/m²), G is Soil heat flux (W/m²) and H is Sensible heat flux (W/m²). This equation represents the residual energy available for evapotranspiration after accounting for soil and sensible heat fluxes.

Evapotranspiration estimation

Finally, Instantaneous Evapotranspiration was derived from latent heat flux (λET) using the following Equation 12⁴⁴:

12

Then, Reference ET Fraction was calculated using Equation 13⁵⁷ where ETo ​ is the reference evapotranspiration from weather data.

13

Finally, thematic maps were generated, processed, and finalized using GIS and remote sensing tools to ensure spatial accuracy and analytical consistency. The flowsheet of methodology is shown in Figure 2.

Fig. 2.

Methodological framework illustrating the integration of Landsat and meteorological data to estimate evapotranspiration using the SEBAL approach.

Results

The comprehensive results and their interpretation generated by assessing the surface energy balance components across Pakistan through SEBAL model revealed that how regional differences in vegetation, land use and climatic parameters influence Albedo, Net Radiation, LST, NDVI, and Evapotranspiration.

LST trends

The spatiotemporal distribution of Land Surface Temperature (LST) across Pakistan for the years 1989, 1999, 2009, and 2019 is illustrated in Figure 3. In terms of LST, the results showed a consistent rise in maximum temperature over time, with the ranges of -20 °C to 45 °C in 1989, -18 °C to 46 °C in 1999, -15 °C to 49 °C in 2009 and -26 °C to 50 °C in 2019. A steady increase in LST was observed over time, with significant warming trends in southern and central regions. Cooler temperatures are consistently recorded in northern high-altitude zones, while urban and arid areas exhibit intensified heat signatures, particularly in 2019.

Fig. 3.

LST Distribution Across Pakistan (1989–2019): Showing warming in southern and central plains and a maximum of 50 °C in 2019.

Vegetation dynamics (NDVI)

NDVI for the period of 1989 to 2019 indicates distinct spatial and temporal changes in vegetation cover across Pakistan ranging from 0.680 to – 0.583 in 1989, 0.865 to -0.424 in 1999, 0.854 to 0.619 in 2009 and 0.903 to -0.679 in 2019 as shown in the Figures 4. Zone A (Northern Highlands) consistently exhibited the highest NDVI values, indicating dense vegetation throughout the period. Zone B (Upper Indus Plains) reflected moderate to high NDVI, with a gradual increase up to 2009, followed by a slight decline by 2019. Zone C (Western Dry Mountains) maintained low NDVI across all years, reflecting its arid nature, with minor greening in 1999 and 2009. Zone D (Lower Indus Plains) reflected an increasing trend until 2009, especially in irrigated areas, but declined again in 2019. Zone E (Southern Arid Zone) consistently reflected the lowest NDVI values, with minimal changes across the years. Results indicated that NDVI values improved between 1989 and 2009 in most zones, while a notable decline was observed in 2019 potentially linked to anthropogenic pressures and climatic extremes. While, EVI data was incorporated within the SEBAL modeling framework to refine vegetation characterization and improve ET estimation accuracy. The integrated EVI imput enhanced the model’s sensitivity to canopy structure, particularly in dense vegetation and irrigated regions. Consequently, areas with higher EVI corresponded to elevated ET across Pakistan climatic zones.

Fig. 4.

Spatiotemporal Distribution of NDVI Across Pakistan (1989–2019): Showing increasing trend until 2009 and a notable decline in 2019.

Surface albedo changes

The results of surface albedo indicate a gradual decline in reflectivity across land surface of Pakistan. A higher albedo indicates land surface reflection of sunlight pointing to regions with higher brightness or reflecting qualities. On the other hand, the lower albedo suggests the surfaces with comparatively less reflection of sunlight as shown in Figure 5. The surface albedo analysis from 1989 to 2019 shows a gradual decline across Pakistan which reflects changing surface and atmospheric conditions over time. In 1989, albedo values ranged from 0.11 to 0.32, with higher reflectance in the northern and arid regions. By 1999, the maximum albedo increased slightly to 0.34. However, from 2009 onward, albedo values decreased, reaching a lower maximum of 0.28 in 2019. This trend reflects a shift from brighter, bare surfaces toward darker, urban or developed areas, indicating changing land surface characteristics over time.

Fig. 5.

Spatiotemporal Variation in Surface Albedo Across Pakistan (1989–2019): Showing a gradual decline, with brighter northern and arid regions shifting to darker, urbanized areas by 2019.

Net radiation variability

The Net Radiation distribution across Pakistan for the years 1989 to 2019 depicts significant spatial variability and temporal increase throughout the study period (Figure 6). In 1989, the value ranged from 190 to615W/m² with moderate cover throughout the country while higher intensities have been observed in Zone C, E and D including Baluchistan, southern Punjab and interior Sindh. However, Zone A along with some parts of Zone B consistently showed lower net radiations due to terrain and high altitude characteristics. In 1999, net radiation ranges slightly expanded to 210 to 635 W/ m² with consistent higher net radiations across Zone C, E and D including central Indus plain and southern arid belts. By 2019, net radiation values show a continued upward trend in the maximum. In 2009 with the values ranged from 205 to 645 W/m², certain southern and south western areas of Zone C exhibited a reduction in comparison to previous decades with an increase in central and eastern Punjab (Zone D and some parts of Zone B) while low values persist in northern areas (Zone A) of Pakistan. In 2019, further increase in the values have been observed ranging from 225 to 650 W/ m² marking the highest values recorded during three decades. The spatial extent of high ranges of net radiations has been expanded across Zone B and some north east areas of Zone D while low values have continuously been observed in Zone A. A distinct comparison between high radiations in Zone B, C, D, E and low radiations in Zone A can easily been observed in 2019 indicating the intensified land surface modifications in Zones. Overall, the temporal findings of net radiations indicate a progressive intensification of maximum net radiation over time with low radiations in northern mountainous zones in comparison to the other zones of Pakistan having high net radiation values.

Fig. 6.

Spatiotemporal Distribution of Net Radiation Across Pakistan (1989–2019): Increasing over time in central and southern zones, while remaining low in northern highlands.

Spatiotemporal analysis of ET

The spatial distribution of evapotranspiration (ET) across Pakistan was assessed using the SEBAL model for four years: 1989, 1999, 2009, and 2019(Figure 7). ET range in different zones represent varying degrees of vegetation cover, moisture availability, and land surface conditions. The ET values range from 4.67 mm/day to 6.0 mm/day that depicts variations in ET intensity from very low to very high. High values close to 6.0 mm/day are observed in Northern areas (Zone A) and in Southern Sindh (Zone E). High ET rate in 1989 has also been observed in Zone B. While low ET values were observed in central and east Punjab (Zone B and D) and in some parts of Sindh (Zone E) and South Western Balochistan (Zone C and E). In 1999, values ranging from 5.0 mm/day to 0.33 mm/day indicating very low to very high respectively. A noticeable expansion can be seen as southern and central Punjab (Zone D) and most of the Balochistan (Zone C and E) show very low ET range while Zone A and Zone B show consistency of high ET range. Moderate ranges of ET have been observed around the study area map during 1999 (Figure 7). In 2009, a temporary resurgence in ET rate was observed across Pakistan including South west parts of Zone D and E. The ET values ranging from 5.33 mm/day to 7.0 mm/day for the year of 2009 as shown in Figure 7. On the contrary, in 2019, a sharp decline in ET, ranging from 5.67 mm/day to 7.33 mm/day, was observed across Pakistan. In the Zone A (northern parts) and Zone B, very high rate of ET has been switched towards high and moderately high while low and very low values have been observed in the zones C, D and E. Overall, ET rates in 2019 illustrates a significant spatial shift from very high and moderate ET zones to low and very low ET zones. Quantitative results with area (in km²) and percentage coverage of each ET class over the four study years are summarized in Table 3. The results highlight a significant decline in areas with very high (6.94 – 7.33), high (6.55 – 6.93) and moderate ET (6.16 – 6.54) over time, while low (5.77 – 6.15) and very low (4.67 – 5.76) categories have expanded, especially by 2019.

Fig. 7.

Spatiotemporal Distribution of ET in Pakistan (1989–2019): showing a decline in high ET areas and expansion of low ET zones by 2019.

Table 3.

Temporal Distribution of Evapotranspiration (ET) Classes in Pakistan from 1989 to 2019

Class	ET Range	1989		1999		2009		2019
Units	(mm/day)	Area(km2)	%	Area(km2)	%	Area (km2)	%	Area (km2)	%
Very high	6.94 – 7.33	147,168	15.0%	90,969	9.2%	140,159	14.2%	51,671	5.2%
High	6.55 – 6.93	61,523	6.2%	103,348	10.5%	62,622	6.3%	97,218	9.8%
Moderate	6.16 – 6.54	88,551	9.0%	165,000	16.7%	68,637	7.0%	66,293	6.7%
Low	5.77 – 6.15	270,978	27.4%	155,398	15.7%	379,674	38.5%	256,368	26.0%
Very low	4.67 – 5.76	418,936	42.4%	472,441	47.9%	336,064	34.0%	515,606	52.3%

Note: “ET classes based on natural breaks (Jenks optimization), consistent with SEBAL standards”⁵.

Discussion

The observed variations reveal a strong interconnection among NDVI, LST and ET, where vegetation dynamics regulate surface temperature and evapotranspiration through energy and moisture exchange⁴⁵. The resulted NDVI trends highlight the combined influence of natural and anthropogenic factors on vegetation dynamics in Pakistan. The general increase in vegetation cover up to 2009 reflects improvements in agricultural practices, irrigation development, and possibly favorable climatic conditions^46,47. However, the decline in NDVI observed after 2009 points to increasing urbanization, land degradation, and water scarcity^48,49. Vegetation remained consistently low in arid and desert regions, with only minor temporary improvements, likely due to localized rainfall or afforestation efforts^50,51. The steady increase of approximately 4.5°C in LST from 1989 to 2019 aligns with previous findings^52–55, providing quantitative evidence of intensified urbanization, vegetation loss and climate variability across Pakistan^56,57. The arid and semi-arid zones, especially in Sindh and southern Punjab, show the highest LST due to sparse vegetation and high solar radiation, similar findings have been described by Arshad⁵⁸ and Qureshi⁵⁹. In contrast, the northern mountainous regions maintain lower LST levels, consistent with their elevation and vegetation density^60,61. Urban centers such as Karachi, Lahore, and Multan demonstrate pronounced LST, aligned with rapid population growth and land cover changes^54,62,63. The variation in net radiation reflects the influence of surface type, topography, and climate³³. Higher values in southern regions are linked to low vegetation and high temperatures, while lower values in the north result from snow cover and high reflectivity^64,65. The decline in net radiation in 2009 could be linked to increased cloud cover, higher surface reflectance (as also shown in albedo data), or satellite data calibration issues⁶⁶ while the rise in 2019 suggests enhanced surface warming. These shifts highlight the importance of net radiation in understanding energy balance, evapotranspiration, and climate change impacts⁶⁷. The declining albedo trend from 1999 to 2019 indicates significant land surface changes in Pakistan^67–69. Higher albedo in earlier years was linked to bare soil and snow-covered areas, while the later decline reflects increased urbanization and land use changes that reduce surface reflectivity. Lower albedo surfaces absorb more solar radiation, contributing to higher surface temperatures^70,71. The spatial and temporal changes in ET rates highlight significant environmental transformations across Pakistan’s major climatic zones⁷². The expansion of low and very low ET areas suggests an overall decline in vegetation health, soil moisture, and water availability, particularly in the last decade^73–75. Northern Highlands (Zone A) which were initially covered with forests and glaciers, showed high ET in 1989 and 2009 (15% and 14.2% respectively). However, by 2019, very high ET fell to 5.2%, indicating possible glacial retreat and land degradation due to warming^34,76. In Punjab Plains the Irrigated Agriculture Belt (Zone B) ET rates were moderate to high throughout the timeline, with the highest Moderate ET recorded in 1999 (16.7%). However, a drop to 6.7% in 2019 points to inefficient irrigation, groundwater depletion, and shifting cropping patterns^77,78. Balochistan Plateau & Western Arid Zones (Zone C) consistently reflected very low ET throughout all four years, peaking at 52.3% in 2019. This confirms the persistent aridity, minimal vegetation, and scant rainfall⁷⁹. Sindh Deltaic Region (Lower Indus Basin) Zone E formerly moderate to high ET zones due to wetland ecosystems, these now show a consistent drop in ET due to seawater intrusion, loss of mangroves, and upstream water diversions⁸⁰. Urban areas marked reduction in ET is evident, especially in peri-urban areas. Urban sprawl, infrastructure expansion, and vegetation loss contribute to increasing very low ET, consistent with global urban heat island trends^10,81,82. Similar declines in ET rates have also been observed in India and Iran using SEBAL due to anthropogenic stressors ^43,28. NDVI-based studies^83,84 have also revealed the decline the vegetation in the same type of regions where ET has dropped sharply. Bastiaanssen⁸⁵ confirmed that long-term reductions in ET correspond with land degradation and desertification risks in arid climates. The observed decline in high and very high ET zones, alongside the expansion of low ET areas, has profound implications for sustainable development in Pakistan. Reduced ET in urbanization landscapes not only highlights the loss of vegetation but also weakens the ecosystem services like evaporative cooling and microclimate regulation, thereby intensifying the UHI effect and posing challenges to sustainable urban development. Observed changes in ET are closely linked to land use and land cover dynamics. Urban expansion, deforestation and land degradation have contributed significantly to ET patterns. The urban expansion in Pakistan has been expanded by 38.02% between 2000 to 2020⁸⁶ and in Lahore the built up areas has been increased by 28% from 1993 to 202⁸⁷. Concurrently, the population growth rates of Pakistan is 2.55% according to Pakistan’s recent 7^th Population and Housing Census (2023)⁸⁸. These metrics provide a clearer attribution of ET changes to anthropogenic drivers, reinforcing the role of land cover and socio-economic factors in shaping regional water and energy fluxes. In the perspective of climate resilience, the long-term ET trends identified in this study highlighted the regions increasingly exposed to droughts, heat stress, and land degradation⁸⁹. These SEBAL derived ET estimates provide spatially detailed insights for Pakistan , support irrigation scheduling, drought monitoring and urban heat mitigation, improving water use efficiency, early warning and public health⁸⁸. For instance, these results can guide water allocation in Punjab plains (Zone B) under the National Water Policy 2018⁹⁰ and highlights mangrove degradation in the Sindh Delta (Zone E) for targeted restoration under the Climate Change Act, 2017⁹¹. Thus, this study emphasizing the need for adaptive measures including urban greenery, water efficient landscaping and sustainable management practices. Moreover, the shifts in ET rate directly reflect the use of water and ground water recharge, making ET monitoring a valuable tool for water resource management⁸. As temperature rises, evapotranspiration intensifies and precipitation becomes less effective for recharge, resulting in declining of groundwater storage^92,93. These applications, align with Pakistan’s National Water Policy and climate adaptation goals, providing actionable insights for sustainable land and water management⁹⁴. Thus, integrating SEBAL based ET assessments into planning and governance frameworks can therefore provide evidence-based guidance to optimize water allocation, strengthen urban climate resilience, and promote sustainable land and resource use in data-scarce environments of Pakistan. In terms of future perspective, a key limitation of this study is the use of a decadal snapshot approach. Although this method effectively captures long term spatiotemporal trends of ET across Pakistan, it may overlook short-term variations such as droughts or anomalous monsoon events. Future studies should incorporate higher-temporal resolution datasets to capture seasonal ET dynamics to understand the impact of climate extremes in Pakistan water and energy balance. SEBAL derived ET estimates have demonstrated reliable performance in similar semi-arid regions, supported by high quality satellite and meteorological data^61,69,89. While in-situ validation is limited, the results provide a robust assessment of spatial and temporal ET patterns across Pakistan, with future studies able to further refine accuracy through ground-based measurements and multi model comparisons.

Recommendations

Incorporating ET and NDVI indices into urban and regional planning can support heat mitigation and climate resilient land management strategies across Pakistan. The findings also suggest the urgent need to integrate SEBAL based ET monitoring into national water management framework to optimize irrigation scheduling and enhance agricultural water efficiency.

Conclusions

This study highlights significant changes in evapotranspiration and surface energy dynamics across Pakistan from 1989 to 2019, marked by a steady rise in land surface temperature, declining vegetation cover, and the expansion of low ET zones. These patterns reflect the increasing impact of urbanization, land degradation, and climate stress. The SEBAL model proved to be a reliable tool for monitoring ET in data-scarce environments. The expansion of very low ET zones by 52.3% in 2019 signifies a critical reduction in evaporative cooling, intensifying the UHI effect in cities. This directly interprets the increased energy demand for cooling and heightened public health risks. Furthermore, the decline in ET in agricultural zones points to s (Zone B and Zone D) severe water stress, threatening food security and necessitating urgent shifts towards water-efficient cropping patterns. Hence, the findings underscore the urgent need for sustainable urban planning, vegetation restoration, improved water management, and adaptive land use strategies to mitigate rising LST and environmental degradation. Integrating SEBAL based monitoring into national policy can ensure long-term ecological stability and climate resilience across Pakistan’s diverse landscape.

Author contributions

1. Aneeza Islam: Conceive original idea, developed methodology, data collection, analysis, writing-original Draft; software processing 2. Syeda Maria Ali: Supervision, conceptualization, assist in data analysis, Writing-manuscript and Editing 3. Eman Rafi Alamery: Project Administration, writing-review and editing, data curation 4. Khadeijah Yahya Faqeih: Assists in review, editing, proofread and referencing, visualization 5. Maha Abdullah Aldhobaihy: Assist in writing review, editing, proofread and referencing 6. Somayah Moshrif Alamri: Data curation, assists in review, editing, proofread and referencing 7. Iftikhar Ali: Conceptualization, co-supervision, investigation, resources, validation.

Funding

This research was funded by Princess Nourah bint Abdulrahman University Researchers Supporting Project number (PNURSP2025R911), Princess Nourah bint Abdulrahman University, Riyadh, Saudia Arabia.

Data availability

Data is provided within the manuscript or supplementary information files.

Declarations

Competing interest

The authors declare no competing interests.