TerraQ: Spatiotemporal Question-Answering on Satellite Image Archives

Konstantinos Plas; Manolis Koubarakis; Sergios-Anestis Kefalidis

arxiv: 2502.04415 · v1 · pith:RLIRJNCAnew · submitted 2025-02-06 · 💻 cs.CV · cs.AI

TerraQ: Spatiotemporal Question-Answering on Satellite Image Archives

Sergios-Anestis Kefalidis , Konstantinos Plas , Manolis Koubarakis This is my paper

Pith reviewed 2026-05-25 08:30 UTC · model grok-4.3

classification 💻 cs.CV cs.AI

keywords TerraQspatiotemporal question answeringsatellite image archivesnatural language processingEarth Observationknowledge basemetadata filtering

0 comments

The pith

TerraQ processes natural language requests to retrieve satellite images matching metadata and knowledge base criteria from archives.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

TerraQ is presented as a spatiotemporal question-answering engine for satellite image archives. It functions as a natural language processing system that handles requests referring to image metadata and entities drawn from a specialized knowledge base. The system supports queries such as requests for images of rivers near ports in France with snow coverage below 20 percent and cloud coverage above 10 percent. The goal is to improve accessibility of Earth Observation data through interfaces resembling digital assistants.

Core claim

TerraQ is a natural language processing system built to process requests for satellite images satisfying certain criteria. The requests can refer to image metadata and entities from a specialized knowledge base.

What carries the argument

TerraQ, the spatiotemporal question-answering engine that maps natural language queries to image selections using metadata attributes and knowledge base entities.

Load-bearing premise

Natural language processing can reliably map complex user queries involving spatiotemporal criteria, metadata attributes, and knowledge base entities to correct image selections without substantial errors or manual intervention.

What would settle it

Executing the example query about rivers near ports in France and observing whether the returned images satisfy the stated snow coverage, cloud coverage, and location constraints.

Figures

Figures reproduced from arXiv: 2502.04415 by Konstantinos Plas, Manolis Koubarakis, Sergios-Anestis Kefalidis.

read the original abstract

TerraQ is a spatiotemporal question-answering engine for satellite image archives. It is a natural language processing system that is built to process requests for satellite images satisfying certain criteria. The requests can refer to image metadata and entities from a specialized knowledge base (e.g., the Emilia-Romagna region). With it, users can make requests like "Give me a hundred images of rivers near ports in France, with less than 20% snow coverage and more than 10% cloud coverage", thus making Earth Observation data more easily accessible, in-line with the current landscape of digital assistants.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

TerraQ is a high-level system description with one example query but no methods, architecture, or results, leaving the core reliability claim untested.

read the letter

TerraQ is presented as a natural language QA engine for satellite image archives that pulls from metadata and a specialized knowledge base. The paper states the purpose and gives one concrete example about rivers near ports in France with snow and cloud coverage limits. That is essentially the full content provided. The new element is the application of QA techniques to Earth observation archives with spatiotemporal and entity-based constraints. It does a clear job of naming a practical access problem in remote sensing and showing the kind of multi-criteria request users might want to make in plain language. The main limitation is the total lack of technical substance. There is no description of the NLP pipeline, how the knowledge base is built or queried, what models or rules handle the mapping, or any test results on accuracy, coverage, or failure cases. Without those pieces the assumption that the system reliably turns complex queries into correct image selections stays unsupported. The citation pattern and any formal elements are not relevant here because almost nothing is developed. This work would mainly interest people building user interfaces for satellite data or applied NLP researchers scanning for domain applications. A reader could note the use case but would gain no technical insight or evidence to build on. I would not send it for peer review. It needs at least a methods section and some evaluation before it can be assessed as a completed contribution.

Referee Report

1 major / 0 minor

Summary. The paper introduces TerraQ as a spatiotemporal question-answering engine for satellite image archives. It is described as an NLP system that processes natural-language requests for images satisfying criteria drawn from image metadata and entities in a specialized knowledge base (e.g., the Emilia-Romagna region), illustrated by the example query requesting images of rivers near ports in France with snow coverage below 20% and cloud coverage above 10%.

Significance. If the system reliably maps complex spatiotemporal, metadata, and KB constraints to correct image retrievals, it would improve accessibility of Earth Observation data in line with digital-assistant trends. The manuscript, however, contains no implementation details, architecture description, evaluation results, or error analysis, so the practical significance cannot be assessed.

major comments (1)

[Abstract] Abstract: the central claim that TerraQ 'processes' such requests to return satisfying images rests on the untested premise that the NLP pipeline can reliably translate combined spatiotemporal, metadata, and KB constraints without substantial errors; no technical description of the pipeline, no test queries, no performance metrics, and no failure-mode analysis are supplied, leaving the claim unsupported.

Simulated Author's Rebuttal

1 responses · 1 unresolved

We thank the referee for the review. The report correctly identifies that the manuscript consists only of a high-level abstract with no implementation details, architecture, test queries, metrics, or error analysis, leaving the central claim unsupported.

read point-by-point responses

Referee: [Abstract] Abstract: the central claim that TerraQ 'processes' such requests to return satisfying images rests on the untested premise that the NLP pipeline can reliably translate combined spatiotemporal, metadata, and KB constraints without substantial errors; no technical description of the pipeline, no test queries, no performance metrics, and no failure-mode analysis are supplied, leaving the claim unsupported.

Authors: We agree with the referee that the abstract asserts TerraQ processes such requests but supplies none of the requested technical content. The manuscript text provided is limited to the abstract itself and contains no pipeline description, examples, or evaluation. We have no additional implementation or results to include. revision: no

standing simulated objections not resolved

Absence of any implementation details, architecture description, test queries, performance metrics, or failure-mode analysis in the manuscript.

Circularity Check

0 steps flagged

No circularity: high-level system description with no derivations or predictions

full rationale

The paper describes TerraQ as an NLP-based spatiotemporal QA engine for satellite archives that handles natural language queries referencing metadata and a knowledge base. No equations, fitted parameters, predictions, or derivation chains are present in the provided abstract or described content. The central claim is a system capability assertion without any self-referential reductions, self-citations used as load-bearing premises, or renamings of known results. This matches the default expectation of no circularity for non-mathematical system papers.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No free parameters, axioms, or invented entities are described in the abstract.

pith-pipeline@v0.9.0 · 5626 in / 992 out tokens · 21304 ms · 2026-05-25T08:30:54.737700+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

20 extracted references · 20 canonical work pages · 3 internal anchors

[1]

GPT-4 Technical Report

OpenAI, “Gpt-4 technical report,” 2024. [Online]. Available: https://arxiv.org/abs/2303.08774

work page internal anchor Pith review Pith/arXiv arXiv 2024
[2]

Llama 2: Open Foundation and Fine-Tuned Chat Models

H. Touvron, L. Martin, K. Stone, P. Albert, A. Almahairi, Y . Babaei, N. Bashlykov, S. Batra, P. Bhargava, S. Bhosale, D. Bikel, L. Blecher, C. Canton-Ferrer, M. Chen, G. Cucurull, D. Esiobu, J. Fernandes, J. Fu, W. Fu, B. Fuller, C. Gao, V . Goswami, N. Goyal, A. Hartshorn, S. Hosseini, R. Hou, H. Inan, M. Kardas, V . Kerkez, M. Khabsa, I. Kloumann, A. K...

work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2307.09288 2023
[3]

Mistral 7B

A. Q. Jiang, A. Sablayrolles, A. Mensch, C. Bamford, D. S. Chaplot, D. de Las Casas, F. Bressand, G. Lengyel, G. Lample, L. Saulnier, L. R. Lavaud, M. Lachaux, P. Stock, T. L. Scao, T. Lavril, T. Wang, T. Lacroix, and W. E. Sayed, “Mistral 7b,” CoRR, vol. abs/2310.06825, 2023. [Online]. Available: https://doi.org/10.48550/arXiv.2310.06825

work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2310.06825 2023
[4]

Core techniques of question answering systems over knowledge bases: a survey,

D. Diefenbach, V . L ´opez, K. D. Singh, and P. Maret, “Core techniques of question answering systems over knowledge bases: a survey,” Knowl. Inf. Syst. , vol. 55, no. 3, pp. 529–569, 2018. [Online]. Available: https://doi.org/10.1007/ s10115-017-1100-y

work page 2018
[5]

Survey on challenges of question answering in the semantic web,

K. H ¨offner, S. Walter, E. Marx, R. Usbeck, J. Lehmann, and A. N. Ngomo, “Survey on challenges of question answering in the semantic web,” Semantic Web, vol. 8, no. 6, pp. 895–920,

work page
[6]

Available: https://doi.org/10.3233/SW-160247

[Online]. Available: https://doi.org/10.3233/SW-160247

work page doi:10.3233/sw-160247
[7]

Retrieval-augmented generation for knowledge-intensive NLP tasks,

P. S. H. Lewis, E. Perez, A. Piktus, F. Petroni, V . Karpukhin, N. Goyal, H. K ¨uttler, M. Lewis, W. Yih, T. Rockt¨aschel, S. Riedel, and D. Kiela, “Retrieval-augmented generation for knowledge-intensive NLP tasks,” in Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems 2020, NeurIPS 2020, Decem...

work page 2020
[8]

The question answering system geoqa2 and a new benchmark for its evaluation,

S. Kefalidis, D. Punjani, E. Tsalapati, K. Plas, M. Pollali, P. Maret, and M. Koubarakis, “The question answering system geoqa2 and a new benchmark for its evaluation,” Int. J. Appl. Earth Obs. Geoinformation , vol. 134, p. 104203, 2024. [Online]. Available: https://doi.org/10.1016/j.jag.2024.104203

work page doi:10.1016/j.jag.2024.104203 2024
[9]

Earthqa: A question answering engine for earth observation data archives *,

D. Punjani, M. Koubarakis, and E. Tsalapati, “Earthqa: A question answering engine for earth observation data archives *,” in IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2023, Pasadena, CA, USA, July 16-21,

work page 2023
[10]

1396–1399

IEEE, 2023, pp. 1396–1399. [Online]. Available: https://doi.org/10.1109/IGARSS52108.2023.10282475

work page doi:10.1109/igarss52108.2023.10282475 2023
[11]

Global oceans and seas, version 1,

F. M. Institute, “Global oceans and seas, version 1,” 2021. [Online]. Available: https://doi.org/10.14284/542

work page doi:10.14284/542 2021
[12]

Extending the Y AGO2 knowledge graph with precise geospatial knowledge,

N. Karalis, G. M. Mandilaras, and M. Koubarakis, “Extending the Y AGO2 knowledge graph with precise geospatial knowledge,” in The Semantic Web - ISWC 2019 - 18th International Semantic Web Conference, Auckland, New Zealand, October 26-30, 2019, Proceedings, Part II, ser. Lecture Notes in Computer Science, C. Ghidini, O. Hartig, M. Maleshkova, V . Sv ´atek...

work page doi:10.1007/978-3-030-30796-7 2019
[13]

OGC GeoSPARQL - A Geographic Query Language for RDF Data,

Matthew Perry and John Herring, “OGC GeoSPARQL - A Geographic Query Language for RDF Data,” Open Geospatial Consortium, OGC Implementation Standard OGC 11-052r4, Sep. 2012. [Online]. Available: http://www.opengis.net/doc/IS/ geosparql/1.0

work page 2012
[14]

Manning, M

C. D. Manning, M. Surdeanu, J. Bauer, J. R. Finkel, S. Bethard, and D. McClosky, “The stanford corenlp natural language processing toolkit,” in Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics, ACL 2014, June 22-27, 2014, Baltimore, MD, USA, System Demonstrations . The Association for Computer Linguistics, 2014, pp. ...

work page doi:10.3115/v1/p14-5010 2014
[15]

From tagme to W AT: a new entity annotator,

F. Piccinno and P. Ferragina, “From tagme to W AT: a new entity annotator,” in ERD’14, Proceedings of the First ACM International Workshop on Entity Recognition & Disambiguation, July 11, 2014, Gold Coast, Queensland, Australia, D. Carmel, M. Chang, E. Gabrilovich, B. P. Hsu, and K. Wang, Eds. ACM, 2014, pp. 55–62. [Online]. Available: https://doi.org/10....

work page doi:10.1145/2633211.2634350 2014
[16]

Heideltime: High quality rule-based extraction and normalization of temporal expressions,

J. Str ¨otgen and M. Gertz, “Heideltime: High quality rule-based extraction and normalization of temporal expressions,” in Proceedings of the 5th International Workshop on Semantic Evaluation, SemEval@ACL 2010, Uppsala University, Uppsala, Sweden, July 15-16, 2010 , K. Erk and C. Strapparava, Eds. The Association for Computer Linguistics, 2010, pp. 321–32...

work page 2010
[17]

Benchmarking geospatial question answering engines using the dataset geoquestions1089,

S. Kefalidis, D. Punjani, E. Tsalapati, K. Plas, M. Pollali, M. Mitsios, M. Tsokanaridou, M. Koubarakis, and P. Maret, “Benchmarking geospatial question answering engines using the dataset geoquestions1089,” in The Semantic Web - ISWC 2023 - 22nd International Semantic Web Conference, Athens, Greece, November 6-10, 2023, Proceedings, Part II, ser. Lecture...

work page doi:10.1007/978-3-031-47243-5 2023
[18]

Next-generation database interfaces: A survey of llm-based text-to-sql,

Z. Hong, Z. Yuan, Q. Zhang, H. Chen, J. Dong, F. Huang, and X. Huang, “Next-generation database interfaces: A survey of llm-based text-to-sql,” CoRR, vol. abs/2406.08426, 2024. [Online]. Available: https://doi.org/10.48550/arXiv.2406.08426

work page doi:10.48550/arxiv.2406.08426 2024
[19]

Three-dimensional geospatial interlinking with jedai-spatial,

M. Papamichalopoulos, G. Papadakis, G. Mandilaras, M. D. Siampou, N. Mamoulis, and M. Koubarakis, “Three-dimensional geospatial interlinking with jedai-spatial,” CoRR, vol. abs/2205.01905, 2022. [Online]. Available: https://doi.org/10.48550/arXiv.2205.01905

work page doi:10.48550/arxiv.2205.01905 2022
[20]

Translating place-related questions to geosparql queries,

E. Hamzei, M. Tomko, and S. Winter, “Translating place-related questions to geosparql queries,” in WWW ’22: The ACM Web Conference 2022, Virtual Event, Lyon, France, April 25 - 29, 2022, F. Laforest, R. Troncy, E. Simperl, D. Agarwal, A. Gionis, I. Herman, and L. M ´edini, Eds. ACM, 2022, pp. 902–911. [Online]. Available: https://doi.org/10.1145/3485447.3511933

work page doi:10.1145/3485447.3511933 2022

[1] [1]

GPT-4 Technical Report

OpenAI, “Gpt-4 technical report,” 2024. [Online]. Available: https://arxiv.org/abs/2303.08774

work page internal anchor Pith review Pith/arXiv arXiv 2024

[2] [2]

Llama 2: Open Foundation and Fine-Tuned Chat Models

H. Touvron, L. Martin, K. Stone, P. Albert, A. Almahairi, Y . Babaei, N. Bashlykov, S. Batra, P. Bhargava, S. Bhosale, D. Bikel, L. Blecher, C. Canton-Ferrer, M. Chen, G. Cucurull, D. Esiobu, J. Fernandes, J. Fu, W. Fu, B. Fuller, C. Gao, V . Goswami, N. Goyal, A. Hartshorn, S. Hosseini, R. Hou, H. Inan, M. Kardas, V . Kerkez, M. Khabsa, I. Kloumann, A. K...

work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2307.09288 2023

[3] [3]

Mistral 7B

A. Q. Jiang, A. Sablayrolles, A. Mensch, C. Bamford, D. S. Chaplot, D. de Las Casas, F. Bressand, G. Lengyel, G. Lample, L. Saulnier, L. R. Lavaud, M. Lachaux, P. Stock, T. L. Scao, T. Lavril, T. Wang, T. Lacroix, and W. E. Sayed, “Mistral 7b,” CoRR, vol. abs/2310.06825, 2023. [Online]. Available: https://doi.org/10.48550/arXiv.2310.06825

work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2310.06825 2023

[4] [4]

Core techniques of question answering systems over knowledge bases: a survey,

D. Diefenbach, V . L ´opez, K. D. Singh, and P. Maret, “Core techniques of question answering systems over knowledge bases: a survey,” Knowl. Inf. Syst. , vol. 55, no. 3, pp. 529–569, 2018. [Online]. Available: https://doi.org/10.1007/ s10115-017-1100-y

work page 2018

[5] [5]

Survey on challenges of question answering in the semantic web,

K. H ¨offner, S. Walter, E. Marx, R. Usbeck, J. Lehmann, and A. N. Ngomo, “Survey on challenges of question answering in the semantic web,” Semantic Web, vol. 8, no. 6, pp. 895–920,

work page

[6] [6]

Available: https://doi.org/10.3233/SW-160247

[Online]. Available: https://doi.org/10.3233/SW-160247

work page doi:10.3233/sw-160247

[7] [7]

Retrieval-augmented generation for knowledge-intensive NLP tasks,

P. S. H. Lewis, E. Perez, A. Piktus, F. Petroni, V . Karpukhin, N. Goyal, H. K ¨uttler, M. Lewis, W. Yih, T. Rockt¨aschel, S. Riedel, and D. Kiela, “Retrieval-augmented generation for knowledge-intensive NLP tasks,” in Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems 2020, NeurIPS 2020, Decem...

work page 2020

[8] [8]

The question answering system geoqa2 and a new benchmark for its evaluation,

S. Kefalidis, D. Punjani, E. Tsalapati, K. Plas, M. Pollali, P. Maret, and M. Koubarakis, “The question answering system geoqa2 and a new benchmark for its evaluation,” Int. J. Appl. Earth Obs. Geoinformation , vol. 134, p. 104203, 2024. [Online]. Available: https://doi.org/10.1016/j.jag.2024.104203

work page doi:10.1016/j.jag.2024.104203 2024

[9] [9]

Earthqa: A question answering engine for earth observation data archives *,

D. Punjani, M. Koubarakis, and E. Tsalapati, “Earthqa: A question answering engine for earth observation data archives *,” in IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2023, Pasadena, CA, USA, July 16-21,

work page 2023

[10] [10]

1396–1399

IEEE, 2023, pp. 1396–1399. [Online]. Available: https://doi.org/10.1109/IGARSS52108.2023.10282475

work page doi:10.1109/igarss52108.2023.10282475 2023

[11] [11]

Global oceans and seas, version 1,

F. M. Institute, “Global oceans and seas, version 1,” 2021. [Online]. Available: https://doi.org/10.14284/542

work page doi:10.14284/542 2021

[12] [12]

Extending the Y AGO2 knowledge graph with precise geospatial knowledge,

N. Karalis, G. M. Mandilaras, and M. Koubarakis, “Extending the Y AGO2 knowledge graph with precise geospatial knowledge,” in The Semantic Web - ISWC 2019 - 18th International Semantic Web Conference, Auckland, New Zealand, October 26-30, 2019, Proceedings, Part II, ser. Lecture Notes in Computer Science, C. Ghidini, O. Hartig, M. Maleshkova, V . Sv ´atek...

work page doi:10.1007/978-3-030-30796-7 2019

[13] [13]

OGC GeoSPARQL - A Geographic Query Language for RDF Data,

Matthew Perry and John Herring, “OGC GeoSPARQL - A Geographic Query Language for RDF Data,” Open Geospatial Consortium, OGC Implementation Standard OGC 11-052r4, Sep. 2012. [Online]. Available: http://www.opengis.net/doc/IS/ geosparql/1.0

work page 2012

[14] [14]

Manning, M

C. D. Manning, M. Surdeanu, J. Bauer, J. R. Finkel, S. Bethard, and D. McClosky, “The stanford corenlp natural language processing toolkit,” in Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics, ACL 2014, June 22-27, 2014, Baltimore, MD, USA, System Demonstrations . The Association for Computer Linguistics, 2014, pp. ...

work page doi:10.3115/v1/p14-5010 2014

[15] [15]

From tagme to W AT: a new entity annotator,

F. Piccinno and P. Ferragina, “From tagme to W AT: a new entity annotator,” in ERD’14, Proceedings of the First ACM International Workshop on Entity Recognition & Disambiguation, July 11, 2014, Gold Coast, Queensland, Australia, D. Carmel, M. Chang, E. Gabrilovich, B. P. Hsu, and K. Wang, Eds. ACM, 2014, pp. 55–62. [Online]. Available: https://doi.org/10....

work page doi:10.1145/2633211.2634350 2014

[16] [16]

Heideltime: High quality rule-based extraction and normalization of temporal expressions,

J. Str ¨otgen and M. Gertz, “Heideltime: High quality rule-based extraction and normalization of temporal expressions,” in Proceedings of the 5th International Workshop on Semantic Evaluation, SemEval@ACL 2010, Uppsala University, Uppsala, Sweden, July 15-16, 2010 , K. Erk and C. Strapparava, Eds. The Association for Computer Linguistics, 2010, pp. 321–32...

work page 2010

[17] [17]

Benchmarking geospatial question answering engines using the dataset geoquestions1089,

S. Kefalidis, D. Punjani, E. Tsalapati, K. Plas, M. Pollali, M. Mitsios, M. Tsokanaridou, M. Koubarakis, and P. Maret, “Benchmarking geospatial question answering engines using the dataset geoquestions1089,” in The Semantic Web - ISWC 2023 - 22nd International Semantic Web Conference, Athens, Greece, November 6-10, 2023, Proceedings, Part II, ser. Lecture...

work page doi:10.1007/978-3-031-47243-5 2023

[18] [18]

Next-generation database interfaces: A survey of llm-based text-to-sql,

Z. Hong, Z. Yuan, Q. Zhang, H. Chen, J. Dong, F. Huang, and X. Huang, “Next-generation database interfaces: A survey of llm-based text-to-sql,” CoRR, vol. abs/2406.08426, 2024. [Online]. Available: https://doi.org/10.48550/arXiv.2406.08426

work page doi:10.48550/arxiv.2406.08426 2024

[19] [19]

Three-dimensional geospatial interlinking with jedai-spatial,

M. Papamichalopoulos, G. Papadakis, G. Mandilaras, M. D. Siampou, N. Mamoulis, and M. Koubarakis, “Three-dimensional geospatial interlinking with jedai-spatial,” CoRR, vol. abs/2205.01905, 2022. [Online]. Available: https://doi.org/10.48550/arXiv.2205.01905

work page doi:10.48550/arxiv.2205.01905 2022

[20] [20]

Translating place-related questions to geosparql queries,

E. Hamzei, M. Tomko, and S. Winter, “Translating place-related questions to geosparql queries,” in WWW ’22: The ACM Web Conference 2022, Virtual Event, Lyon, France, April 25 - 29, 2022, F. Laforest, R. Troncy, E. Simperl, D. Agarwal, A. Gionis, I. Herman, and L. M ´edini, Eds. ACM, 2022, pp. 902–911. [Online]. Available: https://doi.org/10.1145/3485447.3511933

work page doi:10.1145/3485447.3511933 2022