SEB-PV Model Dataset for Evapotranspiration Estimation in Hazelnut and Pistachio Orchards using Gridded Soil Moisture and Temporal Upscaling Methods

SEB-PV Model Dataset for Evapotranspiration Estimation in Hazelnut and Pistachio Orchards using Gridded Soil Moisture and Temporal Upscaling Methods

Description

This dataset contains model inputs and outputs from the Surface Energy Balance for Partially Vegetated surfaces (SEB-PV) model, generated for the research article: "Enhancing Evapotranspiration Estimates in Orchards with the Surface Energy Balance for Partially Vegetated surfaces (SEB-PV) Model through Combined Use of Gridded Soil Moisture and Temporal Upscaling Methods" submitted in Science of the Total Environment.

The dataset supports two primary research objectives:

Evaluate the operational applicability of the SEB-PV model by assessing the substitution of field-measured soil moisture with globally available gridded products (SMAP L4 and CFSv2)

Provide evidence-based temporal upscaling by comparing seven different algorithms for converting instantaneous evapotranspiration (ET) to daily values

Important Note: This dataset contains SEB-PV model inputs and outputs generated for different model configurations. The dataset includes:

Model inputs: Processed field measurements and satellite data (air temperature, relative humidity, soil moisture, Landsat spectral bands, vegetation indices)

Model outputs: Energy balance components (ET, H, G, RN) and daily ET estimates from seven temporal upscaling methods (ETd1-ETd5)

Model parameters: Calibrated coefficients (beta, c1) optimized for each configuration

The SEB-PV model was calibrated and validated against independent flux tower measurements. These files represent the complete modeling dataset under different input source configurations, enabling replication of the study's analyses and further research applications.

Study Sites and Data Coverage

Hazelnut Orchards (Chile)

Larqui: 36°42'30.53"S, 72°21'38.25"W, 72m ASL

Variety: Lewis (planted 2013)

Spacing: 5.0 × 3.5 m

Climate: Warm-summer Mediterranean (Csb)

Data period: 2017-2022

Pullami: 36°35'20.07"S, 71°47'55.51"W, 254m ASL

Variety: Tonda di Giffoni (planted 2011)

Spacing: 5.0 × 2.5 m

Climate: Warm-summer Mediterranean (Csb)

Data period: 2017-2022

Pistachio Orchards (California, USA)

Flores: 36°14'34.46"N, 119°56'16.44"W, 72m ASL

Variety: Kerman on Pioneer Gold 1 rootstock (30-year-old)

Spacing: 5.0 × 5.0 m

Climate: Cold semi-arid (BSk)

Data period: 2015-2019

Nichols: 36°15'29.25"N, 119°30'27.04"W, 77m ASL

Variety: Kerman on Pioneer Gold 1 rootstock (planted 1985)

Spacing: 6.0 × 5.0 m

Climate: Cold semi-arid (BSk)

Data period: 2015-2019

Dataset Structure

File Organization

The dataset contains 20 CSV files following the naming convention: [Site]_[TimeStep]_[SoilMoistureSource]_[Version]_[Beta].csv

Where:

Site: Flores, Larqui, Nichols, Pullami

TimeStep:

hlyinst: hourly instantaneous (agro-meteorological station data)

hhlyinst: half-hourly instantaneous (flux tower meteorological data)

SoilMoistureSource:

measured_defbeta: Field measurements from in-situ sensors

CFSv2_defbeta5cm: CFSv2 gridded product (0.25° resolution, 5cm depth)

SMAP_defbetasurface: SMAP L4 gridded product (9 km resolution, 0-5cm depth)

Version: Nuevo (latest model version)

Beta: Optimal beta parameter value from model calibration

Key Variables

Model Inputs

Meteorological Variables:

Ta: Air temperature (°C)

RH: Relative humidity (%)

VWC: Volumetric water content in-row (m³ m⁻³)

VWC_1: Volumetric water content between-row (m³ m⁻³)

Satellite-Derived Variables:

LST_C: Land surface temperature from Landsat (°C)

LAI: Leaf area index (m² m⁻²)

NDVI: Normalized Difference Vegetation Index

Pv: Vegetation fraction cover

SR_B1-SR_B7: Landsat surface reflectance bands

ST_B10: Landsat surface temperature band

Model Outputs

Instantaneous Energy Balance Components (W m⁻²):

ET: Latent heat flux

H: Sensible heat flux

G: Soil heat flux

RN: Net radiation

Daily ET Estimates (mm day⁻¹): Seven temporal upscaling methods tested:

Group 1 (Requiring measured net radiation):

ETd1: Constant evaporative fraction with measured available energy (Rn-G)

ETd2med: Constant evaporative fraction with measured Rn (daily G=0)

ETd3med: Ratio of instantaneous to daily measured net radiation

Group 2 (Using meteorological variables):

ETd2: Constant evaporative fraction with estimated Rn (daily G=0)

ETd3: Ratio of instantaneous to daily estimated net radiation

ETd4: Ratio of instantaneous to daily solar radiation

ETd5: Ratio of instantaneous to daily reference ET (ET₀)

Model Parameters

beta: Calibrated soil surface resistance parameter (dimensionless)

c1: Calibrated canopy resistance coefficient (dimensionless)

Additional Variables

The files contain additional model inputs and intermediate variables as defined in Lagos et al. (2012), including resistances (rc, rs, rs2), roughness parameters (d, zoo), and atmospheric parameters.

Methodology

Model Calibration

Calibration period: 2 years per site (when data available)

Validation period: Independent years for model evaluation

Optimization metric: Model Decision Making Indicator (MDMI) combining Kling-Gupta efficiency (KGE) and normalized root-mean-square error (NRMSE)

Satellite data: Landsat 8/9 Collection 2 Level-2 products

Flux validation: Eddy covariance measurements with 90% footprint contribution areas

Input Data Sources

Meteorological data:

Chile: INIA agro-meteorological network

California: CIMIS network

Soil moisture products:

SMAP Level-4: 9-km resolution, 3-hourly

CFSv2: 0.25° resolution, 6-hourly

Soil properties: SoilGrids250m 2.0

Model Performance Summary

Validation Results (MDMI values)

Hazelnut orchards: >70 (NRMSE ~21%)

Pistachio orchards: >57 (NRMSE ~29%)

Optimal Temporal Upscaling Methods

Mediterranean hazelnut orchards: ETd2med (RMSE ~0.8 mm day⁻¹)

Semi-arid pistachio orchards: ETd4 (RMSE ~1.3 mm day⁻¹)

Applications and Use Cases

This dataset enables researchers to:

Validate model performance metrics reported in the study using the complete SEB-PV model outputs across different configurations

Test alternative temporal upscaling algorithms by applying new methods to the provided instantaneous model outputs and comparing with the seven methods included (ETd1-ETd5)

Analyze model sensitivity to different soil moisture input sources (measured vs. CFSv2 vs. SMAP L4) and parameter configurations using the provided model runs

Compare SEB-PV performance with other surface energy balance models by using the same processed inputs (meteorological data, satellite products, soil properties) as model forcing

Develop improved parameterization schemes for orchard systems using the calibrated parameters (beta, c1) across different sites and input configurations

Extend research to similar orchard cropping systems by applying the demonstrated modeling framework and parameter ranges to new sites

Study energy balance partitioning in partially vegetated surfaces using the complete energy flux estimates (ET, H, G, RN) under different modeling scenarios

Evaluate operational feasibility of using globally available datasets (gridded soil moisture, weather stations) versus research-grade measurements for ET estimation

Benchmark new evapotranspiration models against the SEB-PV results using identical input datasets and evaluation periods

Investigate climate and crop-specific patterns in temporal upscaling performance by analyzing the provided daily ET estimates across Mediterranean and semi-arid conditions

Data Quality and Limitations

Strengths

Multi-year coverage across contrasting climatic conditions (Mediterranean vs. semi-arid)

Multiple data source evaluation comparing operational vs. research-grade inputs

Standardized modeling framework applied consistently across all site and configuration combinations

Parameter optimization using novel Model Decision Making Indicator (MDMI)

Complete model dataset including inputs, outputs, and calibrated parameters for reproducible research

Operational feasibility assessment demonstrating performance with globally available datasets

Limitations

Instantaneous temporal resolution limited to satellite overpass times (11:30-12:00 local time)

Model framework dependency requiring understanding of SEB-PV structure for proper interpretation

Site-specific parameter calibration with sensitivity varying by location and input configuration

Spatial resolution constraints from gridded soil moisture products (9 km for SMAP L4, 0.25° for CFSv2)

Limited geographic scope covering only Mediterranean and semi-arid orchard systems

Cloud-free image requirement resulting in irregular temporal sampling based on satellite data availability

Related Datasets

Input data sources are available from:

INIA Chile: https://agrometeorologia.cl

CIMIS California: https://cimis.water.ca.gov

SMAP L4: https://doi.org/10.5067/EVKPQZ4AFC4D

CFSv2: https://cfs.ncep.noaa.gov

SoilGrids: https://soilgrids.org

Citation

If you use this dataset, please cite:

Dataset Citation: Cigarra Guíñez, L.E., Lagos, O., Steduto, P., Krogh, S.A., Shapiro, K., Souto, C., Lillo-Saavedra, M., Balbontín, C., Zaccaria, D. (2025). SEB-PV Model Dataset for Evapotranspiration Estimation in Hazelnut and Pistachio Orchards using Gridded Soil Moisture and Temporal Upscaling Methods. Zenodo. https://doi.org/10.5281/zenodo.15476997

Funding

This research was supported by:

ANID-Subdirección de Capital Humano/Doctorado Nacional/2022-21221319

Chileflux scientific network (ANID/FSEQ210019)

Consorcio Tecnológico del Agua CoTH2O (CORFO/20CTECGH145896)

Centro de Recursos Hídricos CRHIAM (ANID/FONDAP/1523A0001)

Acknowledgments

We thank Richard Snyder, Frank Anderson, Carlos Cea, and Samuel G. Metcalf for their assistance with flux tower data collection and field measurements.

Version: 1.0Last Updated: 2025Format: CSVTotal Files: 20License: CC BY 4.0