{"id":"257a21c3-62ec-4ee1-8480-58b877b8d2c9","arxiv_id":"2411.13680","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":2.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A narrative review of long-term mosquito-borne disease forecasting models concludes temperature is the strongest climate predictor, precipitation has mixed effects, and socio-demographic data remain underused.","lead":"This paper is a narrative review of mathematical and statistical models that forecast mosquito-borne disease incidence over long horizons. It finds temperature is the dominant climate driver in the reviewed models, while rainfall has a more subtle role and socio-demographic variables are often left out.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claim that temperature has 'considerable incidence predictive capacity' is unsupported: most reviewed studies report scenario projections or causal analyses, not validated incidence forecasts.","rationale":"The reader flagged sample representativeness as the weakest premise. I agree that the absence of screening counts is a methodological limitation, but the more load-bearing problem is internal: even within the papers the authors chose, the evidence does not support the abstract's quantitative-sounding claim of 'considerable incidence predictive capacity.' A review can be representative and still overstate what the included studies show. The abstract asserts a strong empirical conclusion, yet the manuscript's own Tables 1 and 2 describe only one long-term study with predictive validation; the others are scenario projections, causal analyses, or models without skill metrics. Consequently, the central claim would fail even if the sample were perfect. The concrete test is low-cost: code the studies as described. Because this is a writing/evidence-calibration issue rather than a fundamental invalidity of the review's scope, the verdict should remain CONDITIONAL (unchanged), pending revision of the abstract and a more careful separation of 'frequently used' from 'demonstrated predictive capacity.'","tokens_in":15168,"tokens_out":5847,"duration_ms":53284,"concrete_test":"Build a table from Sections 3.1 and 3.2 (and Tables 1–2) coding each study cited for the temperature claim on three binary items: (1) produces an actual incidence forecast (not suitability/EP/R0 projection); (2) evaluates forecast skill on data (MAE, correlation, CRPS, AUC, etc.); (3) includes a comparison or ablation isolating temperature's contribution. Count studies with all three 'yes'. If fewer than two pass, revise the abstract sentence from 'considerable incidence predictive capacity' to 'frequently used in reviewed models', and soften 'critical driver' to 'commonly emphasized'. This check is deterministic from the manuscript's own descriptions and settles whether the claim is evidence-based.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim in the abstract—'temperature is a critical driver ... with considerable incidence predictive capacity'—goes beyond what the reviewed studies show, even within the authors' selected sample. Section 3.1 describes six long-term studies. Patz et al. (18) computes epidemic potential from VC and GCM temperature, with no incidence forecast or validation. Ryan et al. (22) maps R0(T) and vector distribution shifts; it does not predict incidence. Muñoz et al. (16) is a causality/time-series analysis, not a predictive model. Van Wyk et al. (26) is a mechanistic projection with fixed birth/death rates and no predictive skill assessment. Petrova et al. (19) does produce long-lead dengue forecasts, but Table 2 lists 'lacks detailed predictive accuracy metrics' as a limitation. Only Colón-González et al. (3) reports predictive skill (MAE, blocked cross-validation), and that model includes population density, mobility, air travel, and GDP alongside temperature. Thus the 'considerable incidence predictive capacity' clause is inferred from the frequency of temperature in Table 3, not from demonstrated predictive performance. Similarly, 'critical driver' is not tested by any comparison or ablation within the reviewed papers: no study shows that removing temperature degrades forecast skill. Because this unsupported inference is the abstract's headline conclusion, the review's main claim is overcalibrated relative to its evidence.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript is a narrative review of long-term predictive models for mosquito-borne diseases (MBDs). The authors describe a literature search in PubMed and Google Scholar, summarize six long-term modeling studies (Patz, Ryan, Muñoz, Colón-González, Van Wyk, Petrova) plus several short-term and interdisciplinary modeling approaches, and tabulate variables, data sources, strengths and limitations. The review's central claims are that temperature is a critical driver in MBD predictive models with considerable incidence predictive capacity, that rainfall has context-dependent effects, and that socio-demographic factors are often neglected and constitute a research opportunity.","tokens_in":15535,"tokens_out":4706,"duration_ms":83260,"significance":"This review is useful as a compact entry point to a heterogeneous literature, and it makes a reasonable high-level observation that temperature-dependent mechanistic ingredients (vectorial capacity, R0(T)) are widely used in long-term MBD projections. The tables (Tables 1-4) collect study objectives, limitations, variables, and data sources in an accessible form, and the discussion of transferable methods from climate science (e.g., multi-model ensembles, Bayesian linear inverse models, analog forecasting) and from structural-demographic theory is constructive. However, the manuscript does not provide original code or machine-checked analyses; its value rests on the accuracy and representativeness of the narrative synthesis. The main claims would be valuable if properly calibrated to the evidence, but they currently overstate the degree to which the reviewed studies demonstrate predictive skill.","major_comments":[{"comment":"The abstract's claim that temperature is 'consistently used in multiple reviewed papers with considerable incidence predictive capacity' is not supported by the studies cited. Patz et al. (18) computes epidemic potential from vectorial capacity and GCM temperature with no incidence forecast or validation. Ryan et al. (22) maps temperature-dependent R0(T) and distribution shifts, not incidence predictions. Muñoz et al. (16) is a causality and time-frequency analysis, not a predictive model. Van Wyk et al. (26) is a mechanistic projection with fixed birth/death rates and no predictive skill assessment. Petrova et al. (19) produces long-lead dengue forecasts, but Table 2 itself lists 'lacks detailed predictive accuracy metrics' as a limitation. Only Colón-González et al. (3) reports MAE and blocked cross-validation, and that model includes population density, mobility, air travel, and GDP alongside temperature. The 'considerable incidence predictive capacity' clause is thus inferred from the frequency of temperature in Table 3 rather than from demonstrated predictive performance; no reviewed study reports an ablation or comparison showing that removing temperature degrades forecast skill. Please rephrase the headline claim to distinguish 'frequently used mechanistic driver' from 'demonstrated predictive capacity', or add a systematic appraisal of predictive validity across the reviewed studies.","section":"Abstract and Section 3.1 (Tables 1-2)"},{"comment":"The inclusion and exclusion criteria are described only in subjective terms and no screening statistics are reported. The text states that articles with 'significant methodological biases or dubious findings' were excluded, and that studies with prediction periods exceeding two years were included 'due to the limited number of studies', but the reader is given no counts of records identified, screened, excluded, or selected, and no flow diagram. Without these numbers, the set of six long-term studies cannot be assessed for representativeness, and the review's conclusion that temperature is the field's critical driver could be an artifact of sample selection. Please provide a PRISMA-style flow chart or at minimum report the numbers of studies retrieved and excluded at each stage, with reasons.","section":"Section 2 (Methodology)"},{"comment":"The review treats the reviewed studies as providing comparable evidence about drivers, but they differ in target quantity (epidemic potential, R0, causality, incidence forecasts), spatial scale, and validation status. The conclusions that temperature is 'critical' and that socio-demographic factors are 'often not considered' require either a systematic comparison of model performance or an explicit statement that these are frequency-based observations about model choices rather than evidence of predictive importance. Please add a limitations paragraph that acknowledges the uneven evidence base and the absence of a systematic risk-of-bias or quality assessment.","section":"Section 7 (Conclusions) and Abstract"}],"minor_comments":[{"comment":"The abstract contains several grammatical errors: 'Thus, is urgent to comprehend' should be 'Thus, it is urgent to comprehend', 'specially' should be 'especially', and 'highlighting the potential facing challenges' is garbled. Please copyedit throughout.","section":"Abstract"},{"comment":"Reference [24] lists 'U. Tatem' as the first author; the systematic review of mathematical models of vector-borne diseases is by Andrew J. Tatem et al. (J R Soc Interface 2013). Please correct the citation.","section":"References, [24]"},{"comment":"The sentence 'The compartmental model, originally conceptualized pioneering byGorgas and Garrison (8) and Macdonald (12), was formally introduced in its seminal form by Kermack and McKendrick (10)' is missing a space and awkwardly credits Kermack-McKendrick after Gorgas-Garrison; please rephrase for clarity and historical accuracy.","section":"Section 3.2"},{"comment":"In the sentence 'Specific variables such as ... where employed by applied, play vital roles', the phrase 'where employed by applied' is unintelligible; please rewrite.","section":"Section 4"},{"comment":"The note that variables are grouped 'arbitrarily' (Section 4) weakens the table; consider providing an explicit definition for each category or a supplementary table with variable definitions and units.","section":"Table 3"},{"comment":"If Figure 1 is included in the final version, ensure the histogram has labeled axes and a legend for category colors; the current text does not describe the figure contents beyond 'Frequency of Usage of Variable Categories in Predictive Models'.","section":"Figure 1"}],"recommendation":"major_revision","confidential_remarks":"The paper is a narrative review and does not claim original modeling results, which is appropriate if the journal's scope includes reviews. My main concern is that the abstract and conclusions overstate predictive capacity relative to what the cited studies actually demonstrate; this is fixable with recalibrated wording and the addition of screening statistics. I would also verify the reference list against the citations, particularly reference [24]."},"author_rebuttal":null,"desk_editor":null,"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["92D30","92-02"],"pacs":[],"model":"deepseek-v4-flash","headline":"Temperature is the critical driver in long-term mosquito-borne disease prediction, and socio-demographic factors are the field's biggest missed opportunity.","keywords":["mosquito-borne diseases","long-term prediction","temperature-dependent transmission","vectorial capacity","basic reproduction number","dengue forecasting","climate change","socio-demographic factors"],"falsifier":"A systematic search with explicit screening counts that locates numerous long-term MBD models whose main predictive variables are precipitation, humidity, or socio-demographic factors, and that perform comparably to temperature-only models, would refute the claim that temperature is the field's consistent critical driver.","tokens_in":14984,"feed_emoji":"🦟","tokens_out":5227,"duration_ms":50664,"temperature":0.7,"pith_summary":"This review asks what actually drives long-term forecasts of mosquito-borne disease incidence, and it answers that temperature is the most reliable and widely used driver. Across the papers surveyed, temperature enters through mechanistic quantities such as vectorial capacity and temperature-dependent basic reproduction number, and it repeatedly delivers meaningful predictive skill. Rainfall is more ambiguous, because moderate rain creates breeding sites while heavy rain flushes larvae. The review also finds that socio-demographic variables such as population density, mobility, and GDP are rarely integrated, and it argues this omission is the field's largest opportunity for more comprehensive long-term forecasts.","feed_headline":"Temperature dominates long-term mosquito-disease forecasts","feed_subtitle":"A narrative review finds climate-driven models lead the field, while socio-demographic factors are rarely included.","key_machinery":"The load-bearing objects are the vectorial capacity equation $VC = \\frac{m b c a^2 p^n}{-\\ln p}$ and the temperature-dependent basic reproduction number $R_0(T)$, built from a Ross-Macdonald-style model with $R_0(T) = \\left[\\frac{a(T)^2 b(T) c(T) \\exp(-PDR(T)/\\mu(T)) \\cdot EFD(T) p_{EA}(T) MDR(T)}{N r \\mu(T)^3}\\right]^{1/2}$. These translate temperature into transmission potential through mosquito biting rate, survival, extrinsic incubation, and egg-to-adult development, and they are what let temperature stand in for the whole transmission cycle in long-term projections. The review uses these quantities as its organizing lens to compare models with prediction horizons beyond two years and to argue that temperature-driven mechanistic structure is the field's current backbone.","core_discovery":"The paper's central claim is that the long-term predictive literature for mosquito-borne diseases is organized around temperature as the critical climate driver, with mechanistic models expressing temperature-dependent mosquito biology as the workhorses of this literature. It documents vectorial capacity simulations, empirically calibrated temperature-dependent $R_0(T)$ functions, and temperature-driven compartmental models as the dominant approaches. Precipitation appears in several models but its effect is nonlinear and region-dependent, so it enters less consistently. The converse finding is an absence: socio-demographic factors such as population density, human mobility, air travel, and GDP appear in only a minority of models, and the review concludes that integrating them is the clearest path toward more reliable long-term forecasts.","pith_inferences":["The temperature-centric synthesis may partly reflect the review's own selection; the authors' inclusion criteria are subjective and no screening counts are given, so a broader corpus might show precipitation-driven or socio-demographic models as more prominent.","A testable extension is to code every long-term MBD model in a systematic corpus for which variable classes appear in the final model, then compare predictive skill of temperature-only versus temperature-plus-demography models; this would show whether the identified gap is a modeling gap or a data gap.","The paper's decadal horizon implies that static assumptions, such as constant human population and fixed birth and death rates, are the first assumptions to relax, because migration and demographic change accumulate over ten years.","Borrowing meteorology's ensemble and bias-correction toolbox points toward replacing point forecasts of MBD incidence with probabilistic multi-scenario forecasts, which is what public-health planning would actually consume."],"forward_implications":["If temperature is the critical driver, long-term MBD forecasts should prioritize accurate decadal temperature projections, including bias-corrected multi-model ensembles from climate science.","Mechanistic quantities such as vectorial capacity and $R_0(T)$ should remain central in next-generation models, because they give climate inputs a biological interpretation.","Rainfall must be modeled nonlinearly: moderate precipitation can amplify transmission while excessive rainfall can suppress it, so simple linear rainfall terms are likely to mislead.","Socio-demographic variables are currently underrepresented, so models that add population density, mobility, GDP, and urbanization have the greatest room to improve long-term forecasts.","Techniques from meteorology and demography, including ensemble averaging, bias correction, Bayesian parameter estimation, and structural-demographic indicators, are transferable to MBD prediction."],"supporting_citations":[{"why":"Introduces the vectorial capacity simulation with GCM temperature forecasts that anchors the review's temperature-centric mechanistic strand.","marker":"(18)"},{"why":"Supplies the empirically calibrated temperature-dependent $R_0(T)$ model used to project Aedes-borne virus risk to 2050 and 2080.","marker":"(22)"},{"why":"Provides the mechanistic temperature-trait framework that Ryan et al. and Martheswaran et al. build on.","marker":"(14)"},{"why":"The main counterexample that includes socio-demographic variables such as GDP, mobility, and air travel and shows they can be integrated into long-term dengue projections.","marker":"(3)"},{"why":"Exemplifies temperature-dependent compartmental modeling for long-term Zika and dengue projections in Brazil.","marker":"(26)"},{"why":"Demonstrates long-lead El Nino forecasting coupled to a statistical dengue model, the review's main example of extending climate forecasts into disease early warning.","marker":"(19)"},{"why":"Supports the rainfall subtlety claim by showing precipitation correlates negatively with dengue in Colombia and that ENSO effects are region-dependent.","marker":"(16)"},{"why":"Defines the basic reproduction number $R_0$, the epidemiological quantity the review uses to frame temperature-dependent transmission potential.","marker":"(5)"}],"fun_headline_variants":["Temperature dominates mosquito-borne disease forecasts","Mosquito disease models overlook socio-demographic factors","Climate-driven models lead, but social data is missing","Mechanistic models rely on temperature, ignore society"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The conclusion that temperature is the critical driver rests on the premise that the small, subjectively selected set of reviewed papers represents the broader long-term mosquito-borne disease prediction literature; if that sample is unrepresentative, the synthesis is an artifact of the selection.","fun_headline_variants_meta":{"raw":{"variants":["Temperature dominates mosquito-borne disease forecasts","Mosquito disease models overlook socio-demographic factors","Climate-driven models lead, but social data is missing","Mechanistic models rely on temperature, ignore society"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000724,"raw_usage":{"total_tokens":3222,"prompt_tokens":895,"completion_tokens":2327,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":511,"completion_tokens_details":{"reasoning_tokens":2269}},"tokens_in":511,"tokens_out":2327,"duration_ms":17235,"temperature":1.0,"reasoning_tokens":2269,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T15:59:12.867729+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A systematic search with explicit screening counts that locates numerous long-term MBD models whose main predictive variables are precipitation, humidity, or socio-demographic factors, and that perform comparably to temperature-only models, would refute the claim that temperature is the field's consistent critical driver.","supporting_citations":[],"review_version":1}