2016
Dhungel, R., Allen, R.G., Trezza, R., & Robison, C.W. (2016). Evapotranspiration between satellite overpasses: methodology and case study in agricultural dominant semi-arid areas. Meteorological Applications, 23(4), 714–730. DOI ↗
Scientific History & Publications
Scientific origin, evolution, and publication record.
This page follows the BAITSSS model: where the model came from, how the research evolved across institutions and applications, and where that work appears in the peer-reviewed literature.
01 · Scientific origin
Between satellite observations
Field processes continue between satellite observations. The BAITSSS model was developed to carry the water and energy balance through those intervals using remote sensing, weather, vegetation information, soil properties, and management assumptions.
02 · Selected peer-reviewed record
Publications
2016
Dhungel, R., Allen, R.G., & Trezza, R. (2016). Improving iterative surface energy balance convergence for remote sensing based flux calculation. Journal of Applied Remote Sensing, 10(2), 026033. DOI ↗
2019
Dhungel, R., Aiken, R., Colaizzi, P.D., Lin, X., O'Brien, D., Baumhardt, R.L., Brauer, D.K., Marek, G.W., & Evett, S.R. (2019). Evaluation of uncalibrated energy balance model (BAITSSS) for estimating evapotranspiration in a semiarid, advective climate. Hydrological Processes, 33(15), 2110–2130. DOI ↗
2019
Dhungel, R., Aiken, R., Colaizzi, P.D., Lin, X., Baumhardt, R.L., Evett, S.R., Brauer, D.K., Marek, G.W., & O'Brien, D. (2019). Increased bias in evapotranspiration modeling due to weather and vegetation indices data sources. Agronomy Journal, 111(3), 1407–1424. DOI ↗
2020
Dhungel, R., Aiken, R., Lin, X., Kenyon, S., Colaizzi, P.D., Luhman, R., Baumhardt, R.L., O'Brien, D., Kutikoff, S., & Brauer, D.K. (2020). Restricted water allocations: Landscape-scale energy balance simulations and adjustments in agricultural water applications. Agricultural Water Management, 227, 105854. DOI ↗
2021
Allen, R.G., Dhungel, R., Dhungana, B., Huntington, J.L., Kilic, A., & Morton, C.G. (2021). Conditioning point and gridded weather data under aridity conditions for calculation of reference evapotranspiration. Agricultural Water Management, 245, 106531. DOI ↗
2021
Dhungel, R., Aiken, R., Evett, S.R., Colaizzi, P.D., Marek, G., Moorhead, J.E., Baumhardt, R.L., Brauer, D.K., Kutikoff, S., & Lin, X. (2021). Energy imbalance and evapotranspiration hysteresis under an advective environment: Evidence from lysimeter, eddy covariance, and energy balance modeling. Geophysical Research Letters, 48, e2020GL091203. DOI ↗
2023
Dhungel, R., Anderson, R.G., French, A.N., Saber, M., Sanchez, C.A., & Scudiero, E. (2023). Assessing evapotranspiration in a lettuce crop with a two-source energy balance model. Irrigation Science, 41(2), 183–196. DOI ↗
2024
Dhungel, R., Anderson, R.G., French, A.N., Skaggs, T.H., Ajami, H., & Wang, D. (2024). Intercomparison of citrus evapotranspiration among eddy covariance, OpenET ensemble models, and the water and energy balance model (BAITSSS). Agricultural Water Management, 304, 109066. DOI ↗
2025
Dhungel, R., Anderson, R.G., French, A.N., Skaggs, T.H., Wang, G., Li, N., & Scudiero, E. (2025). Impact of soil heterogeneity on evapotranspiration and irrigation modeling using POLARIS and SSURGO-based soil hydraulic properties. Agricultural Water Management, 319, 109806. DOI ↗
Selected peer-reviewed publications supporting the BAITSSS scientific lineage, model development, forcing methods, and published evaluations. Related conference papers and collaborator publications are not included here.
Selected studies
This visual set follows major stages of the BAITSSS model record: development between satellite overpasses, independent ET evaluation, landscape-scale water-allocation analysis, and citrus ET intercomparison.
Satellite
overpass
overpass
BAITSSS model carries water and energy balance between observations
Next
overpass
overpass
2016 · Method development
Evapotranspiration between satellite overpasses
Question: How can ET be represented through the interval between satellite observation dates while retaining field-scale spatial information?
The paper introduced the BAITSSS model framework for carrying surface energy and soil-water states between satellite overpasses.
Open DOI →Weather +
vegetation inputs
vegetation inputs
BAITSSS model
Weighing lysimeter +
surface measurements
surface measurements
2019 · Independent evaluation
Uncalibrated ET evaluation in a semiarid, advective climate
Question: How does the BAITSSS model perform without prior site calibration when compared with independent ET and surface measurements under strongly advective conditions?
The study used a 127-day corn growing season and compared simulated ET with weighing-lysimeter observations.
Published result snapshot
7%cumulative ET error in the blind test
1.70daily RMSE, mm d⁻¹
0.66daily r²
After revisions described in the paper: cumulative ET error <1%, daily RMSE 1.09 mm d⁻¹, and daily r² 0.86.
Values apply to this study and configuration, not as a universal BAITSSS model accuracy claim.
2020 · Water management application
Restricted water allocations and landscape-scale simulation
Question: How can landscape-scale energy-balance simulation be used to examine changes in agricultural water application under restricted allocations?
The study covered two groundwater regions during 2013–2017 and compared simulated irrigation with reported irrigation within water-right management units.
Open DOI →
2024 · Multi-source intercomparison
Citrus evapotranspiration intercomparison
Question: How do BAITSSS model and OpenET ET estimates compare with eddy covariance observations and water received in citrus orchards?
The study evaluated navel and mandarin orchards over 2021–2022. BAITSSS model ET results were generated using actual irrigation records rather than model-simulated irrigation.
Two-year mean annual cumulative ET
EC 908 mm · OpenET ensemble 1169 mm · BAITSSS model 867 mm · water received 886 mm. The paper reports that BAITSSS model totals were within about ±10% of water received for all site-year combinations; performance against EC varied by orchard and year.
Published measurement comparisons
A 2020 ASABE irrigation-symposium paper summarized field comparisons and irrigation-management experiments from the BAITSSS model research program. The figures below redraw reported values rather than reproduce the original journal or conference artwork.
Bushland 2016 · Corn
Measured versus simulated variables
The published synthesis reports comparisons between 22 May and 26 September 2016 under advective field conditions.
Daily ET
0.85 mmRMSE · r² 0.90Hourly ET
0.10 mmRMSE · r² 0.90Surface temperature
3.27 °CRMSE · r² 0.86Net radiation
48.7 W m⁻²RMSE · r² 0.99Source: BAITSSS Model: An Opportunity to Integrate Remote Sensing and Energy Balance Modeling for In-Season Crop Water Management, ASABE Irrigation Symposium paper 20-065.
Open DOI →SD-6 LEMA · Kansas · 2012
Irrigation-management scenarios
A representative 30 m pixel was used to examine irrigation-system and MAD settings for a corn growing season from 10 May to 15 September.
MAD 0.5 · sprinkler
520 mm
MAD 0.5 · subsurface drip
440 mm
MAD 0.7 · sprinkler
400 mm
Reported irrigation events: 13, 11, and 10 respectively. Relative to MAD 0.5 sprinkler, seasonal irrigation decreased about 15% with subsurface drip and about 23% with MAD 0.7 sprinkler. The paper reports seasonal ET about 11% lower in both alternative scenarios.
Open DOI →Spatial irrigation pattern
The same published synthesis shows a center-pivot example in SD-6 LEMA where seasonal vegetation patterns corresponded with three broad categories of simulated cumulative irrigation.
Published spatial example · 2013
Three irrigation sectors visible within one center pivot
80–160 mmlower cumulative irrigation sector
240–320 mmintermediate cumulative irrigation sector
400–480 mmhigher cumulative irrigation sector
The published discussion relates these spatial differences primarily to seasonal and pixel-scale vegetation variation, with additional influence from soil water-holding characteristics. The authors also note that these simulated spatial patterns could not be fully verified against actual within-pivot irrigation conditions.
This website summary preserves that limitation rather than presenting the simulated sectors as measured irrigation.
Open DOI →These visuals are purpose-built from published numerical values and study descriptions. They are not screenshots of journal figures and they do not replace the publications.
Soil heterogeneity and irrigation response
The BAITSSS model can be used to examine how different soil representations influence soil-water behavior, evapotranspiration components, and modeled irrigation at the same location.
Soil dataset comparison
Soil water and evapotranspiration
The vegetation trajectory is held in common while the BAITSSS model is evaluated with POLARIS and SSURGO soil representations.
Irrigation response
Modeled irrigation comparison
Different soil representations produce different modeled irrigation timing and frequency.
Interpretation
Soil representation changes modeled water balance behavior
Together, the figures show that differences in soil representation can propagate into different root-zone water behavior and irrigation response while the vegetation trajectory remains the same.
Documented research geography
The map shows only numbered study markers. The matching index below carries the study description, coordinate information, and DOI so the geography stays clean and readable.
1
2
3
4
5
6
Documented study area
Exact documented point
State-base map: public-domain U.S. SVG
Numbered markers match the index below.
Numbered markers match the index below.
Study index
1
2016
Near American Falls, Idaho
Published BAITSSS model study area near American Falls, Idaho; shown as an area because the paper does not provide one exact site coordinate.
Open DOI →2
2016
San Joaquin County, California
Published BAITSSS model study area in San Joaquin County, California; shown at regional scale rather than as an invented point.
Open DOI →3
2024
Near Visalia, California
Two documented commercial citrus study sites near Visalia, California, approximately 15 km apart.
Open DOI →4
2023 / 2025
Yuma Valley, Arizona / Bard, California
Documented BAITSSS model study locations in the Yuma Valley and Bard agricultural region.
Open DOI →5
2020+
Northwest Kansas / Sheridan 6
Documented BAITSSS model research area in northwest Kansas, including the Sheridan 6 region.
Open DOI →6
2019 / 2021
Bushland, Texas
Documented USDA-ARS field site at Bushland, Texas, with the published coordinate shown below.
35.186714° N, 102.094189° W · USDA-ARS lysimeter site
Open DOI →Research through time
The timeline records when documented BAITSSS model study locations entered the public research record. Scientific findings remain in the published studies above and in the cited papers.
2016
Idaho and California
Methodology study areas near American Falls and in San Joaquin County.
2019
Bushland, Texas
Documented field-evaluation site.
2020
Northwest Kansas
Documented landscape-scale application region.
2023–2025
Yuma and Bard
Documented agricultural study locations in Arizona and California.
2024
Visalia, California
Documented commercial citrus study sites.
01 · Evidence map
What each evidence category answers
This separation is central to interpreting BAITSSS model results correctly.
| Evidence category | Question answered | Primary authority | Public record |
|---|---|---|---|
| Peer-reviewed scientific evaluation | How did the model perform under the sites, observations, periods, and methods described in the study? | Published measurements, comparison datasets, methods, and peer-reviewed analysis | Published |
| Numerical equivalence | Does a changed execution route reproduce the accepted scientific reference within defined numerical criteria? | Accepted reference implementation and controlled test cases | Software verification |
| Checkpoint and lifecycle testing | Can a run stop, resume, restart, and preserve intended execution state? | Controlled software tests and run records | Software verification |
| Performance benchmarking | How long does a defined computation take under a stated hardware and workload configuration? | Measured execution time for specified test cases | Engineering evidence |
| Independent scientific comparison | How do BAITSSS model outputs compare with independent observations or products at the relevant spatial and temporal scale? | Field measurements, thermal observations, ET products, soil-water observations, and irrigation records | Ongoing evaluation |
02 · Peer-reviewed record
Published scientific evaluation
Published studies
Field and site evaluations
The BAITSSS model has a peer-reviewed publication record covering defined study sites, climates, crops, observations, input sources, and experimental conditions.
The publication itself is the authority for the numerical result, study design, comparison method, and stated limitation.
View publications →Interpretation boundary
No single universal accuracy number
A result from one site or period should not be generalized automatically to every field. Scientific performance must be interpreted in the conditions under which the evaluation was performed.
Accuracy and interpretation FAQ →03 · Software verification
Tests of the BAITSSS Desktop execution system
These tests address whether BAITSSS Desktop executes the intended BAITSSS model scientific workflow reliably. They are not substitutes for field validation.
01
Numerical equivalence
BAITSSS Desktop execution methods are compared with the accepted BAITSSS model scientific reference under controlled conditions and defined numerical criteria.
02
Checkpoint and restart
Stop, resume, restart, and run-lifecycle behavior are tested separately so continuity of scientific state is not assumed from interface behavior.
03
Run identity and provenance
BAITSSS Desktop project inputs, run configuration, software state, outputs, and execution records are preserved so a result can be traced back to the run that produced it.
04 · Computational evidence
Performance is reported as a workload measurement
Benchmarking rule
Workload, hardware, and timing belong together
Runtime is meaningful only with the size of the simulated area, number of model hours, execution configuration, and computer used for the test.
Current performance observations →Interpretation boundary
Speed is not scientific accuracy
A faster execution route is useful only if the scientific state and outputs remain consistent with the accepted reference under the defined verification criteria.
05 · Ongoing scientific comparison
Independent observations and model behavior
Independent observations
Comparison across scale and time
Thermal observations, ET products, field measurements, soil-water observations, and irrigation records can provide independent comparisons. Spatial scale, temporal scale, measurement uncertainty, and management conditions must be considered before interpreting differences.
Model behavior
Sensitivity, resistance, and transferability
Resistance connectivity, parameter sensitivity, stability, assumptions, and transferability remain subjects of scientific evaluation rather than being treated as fixed proof of performance everywhere.
Evidence standard: published science, software verification, computational benchmarking, and ongoing experiments remain separate categories. A result should be traceable to its test conditions and evidence source before it is used as a public claim.