Pith. sign in

REVIEW 3 major objections 8 minor 77 references

Drought risk in Italy is driven more by how the index and its duration memory are modeled than by which evapotranspiration formula is used; southern regions trap long droughts.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.5

2026-07-10 20:56 UTC pith:XNTYLYEH

load-bearing objection Solid applied paper: BATs wins the SPEI distribution bake-off over Italy, and a spatial GP extension of BMCD maps a clear south–north persistence gradient that duration-blind indices miss. the 3 major comments →

arxiv 2607.06805 v1 pith:XNTYLYEH submitted 2026-07-07 stat.AP

A spatial duration-augmented framework for drought persistence

Touqeer Ahmad , Thordis Linda Thorarinsdottir This is my paper
classification stat.AP
keywords SPEIbulk-and-tails distributiondrought persistenceBayesian spatial modelingduration-augmented Markov chainMediterranean droughtreference evapotranspiration
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

This paper argues that standard drought indices alone miss how long dry spells actually last, and that the way those indices are built and the way persistence is modeled matter more than the choice of reference-evapotranspiration formula. Over a 0.25-degree grid of Italy the authors show that three common RET methods produce nearly identical SPEI series once a flexible bulk-and-tails distribution is used to standardize the water-balance deficit. That BATs distribution passes normality checks at 98–99 percent of grid cells, far above the usual GEV, logistic, or normal candidates. They then embed a duration-dependent recovery hazard inside a Bayesian spatial model so that the chance a drought ends can fall as the spell lengthens, and so that this behavior can vary smoothly across the map. The fitted fields reveal a sharp south–north contrast: Sicily, Sardinia and the southern peninsula show strongly declining recovery hazards and heavy survival tails, while the Alps and Po Valley are nearly memoryless. The practical claim is that duration-blind classifications systematically understate multi-year drought risk where the climate itself traps dry conditions.

Core claim

Drought characterization across Italy is far more sensitive to the probability distribution used for SPEI and to explicit duration-dependent persistence than to the RET formulation; once both are modeled, southern peninsular and island regions exhibit pronounced duration-dependent recovery hazards that conventional memoryless classifications understate.

What carries the argument

A Bayesian spatial duration-augmented Markov chain in which the logit recovery hazard is linear in log(duration+1) with spatially varying intercept and slope governed by Matérn Gaussian processes; this yields continuous, uncertainty-aware maps of the total slope β_total(s) and of recovery and survival curves.

Load-bearing premise

The recovery hazard is forced to be linear in log(duration+1) with only a spatially varying intercept and slope; if true duration dependence is non-monotonic or saturating, the southern persistence maps would be an artifact of that form.

What would settle it

Re-fit the spatial model with a non-parametric or higher-order duration effect (for example a spline or step-function hazard) on the same Italian SPEI fields; if the south–north gradient in survival tails disappears or reverses, the claimed duration-trap mechanism fails.

Watch this falsifier — get emailed when new claim-graph text bears on it.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 8 minor

Summary. The paper develops a Bayesian spatial duration-augmented Markov framework for drought persistence over Italy, while jointly assessing how SPEI construction depends on (i) three RET formulations (Penman–Monteith, Hargreaves, Thornthwaite) and (ii) the distribution fitted to the climatic water balance. Using MADIA reanalysis on 730 grids (1981–2022), it reports that spatially averaged SPEI-3/6/12 series are nearly identical across RET methods, whereas the bulk-and-tails (BATs) distribution yields Shapiro–Wilk (and AD/Lilliefors) normality acceptance rates of ~98–100% versus substantially lower rates for GEV, Pe-III, GenLog, and Normal. A spatial GP extension of a duration-augmented binary Markov chain then outperforms a nonspatial counterpart (LOOIC/WAIC; residual Moran’s I near zero), producing continuous recovery/survival fields that show a strong south–north gradient: southern peninsular and island regions exhibit declining recovery hazards with duration, while northern/Alpine regimes are near memoryless. The central claim is that drought characterization is more sensitive to distributional and persistence assumptions than to RET choice, and that duration-insensitive SPEI thresholds understate extended southern drought risk.

Significance. If the empirical ranking holds, the work is a solid, transferable contribution to applied drought statistics: it jointly stress-tests two under-examined SPEI ingredients over a heterogeneous domain, introduces BATs as a competitive SPEI candidate with dual-tail flexibility, and embeds duration-dependent recovery in a spatial Bayesian model that yields uncertainty-aware persistence fields rather than pointwise Hurst/DFA maps alone. Strengths include multi-layered external checks (normality of SPEI, LOO predictive density, residual spatial diagnostics, independent long-memory maps), public R code for BATs fitting, and a clear practical implication for Mediterranean monitoring—that duration-blind classifications can understate multi-year risk in the south. The modular design (RET × distribution × spatial BMCD) is reusable beyond Italy. The result is incremental rather than paradigm-shifting, but methodologically coherent and policy-relevant for a climate-change hotspot.

major comments (3)
  1. §2.5.1, Eq. (8): The recovery hazard is restricted to a logit-linear form in log(D_t(s)+1) with additive spatial intercept and slope. This is the load-bearing parametric choice that generates the mapped β_total field and the heavy southern survival tails in Fig. 10(b–c). The qualitative south–north contrast is corroborated by independent Hurst/DFA maps (Fig. 9), so the ranking of sensitivities is not solely an artifact of Eq. (8); however, quantitative tail mass (e.g., P(duration > 12 months)) is form-dependent. A brief robustness check—duration-category dummies, a nonparametric smooth of duration, or a saturating alternative—or an explicit limitation stating that non-monotonic/saturating hazards are outside the model class would make the survival-curve claims more defensible without changing the paper’s comparative message.
  2. §3.3–3.3.2 and Table 2: The spatial persistence analysis is reported for a single SPEI construction (BATs under RET_PM). The paper’s own thesis is that distributional choice matters more than RET, and that persistence estimates depend on the underlying index. At minimum, confirm that the sign pattern of β_total(s) (negative south, near-zero north) is stable under one alternative distribution (e.g., GEV) and/or one alternative RET at a fixed scale (e.g., SPEI-6). Without that check, the claim that duration-insensitive classifications “systematically understate” southern risk is tied to one pipeline even though Fig. 7 only shows national-mean SPEI agreement across RET.
  3. §2.5, drought indicator definition: Drought is fixed as SPEI_j < −1 with no threshold sensitivity. Because exit probabilities and spell lengths depend on this cut, a short appendix table (e.g., −0.5 / −1 / −1.5 at SPEI-6) showing that the south–north ordering of β_total is preserved would close a natural robustness gap. This is secondary to the two points above but is standard for binary drought-state models.
minor comments (8)
  1. Abstract and elsewhere: “Hargraves” should be “Hargreaves”; “identical” in the abstract is stronger than “nearly identical” used in the body (Fig. 7).
  2. Notation inconsistency: RET methods appear as PM/PEM, TW, HG; Fig. 5 caption and some results text use “PEM” while equations use PM. Standardize to PM throughout.
  3. §2.2 Eq. (1) text: typo “refenece evapotranspiration”; also “Hargraves-based” in the same subsection.
  4. Fig. 7 legend: “PEM” again; axis label “Y ear” has a space typo.
  5. §2.5.1: Prior mean for range parameters is set to 0.25×L; state the numerical value of L (or the resulting prior means for ρ_α, ρ_β) for reproducibility.
  6. §3.2 and Supplementary: Grid-wise SW maps for BATs (Fig. 6) are clear; consider adding a one-sentence note that multiple-testing adjustment is Benjamini–Hochberg across grids×scales so readers do not misread raw p-value maps.
  7. References: Doizé et al. (2026) and related 2025–2026 items are fine as preprints, but ensure arXiv IDs are listed where available for permanence.
  8. Introduction/formatting: Several word-run-ons appear to be PDF extraction artifacts (“Droughtmonitoringacross…”, “inducingcumulativedamage…”); clean the source so the published PDF does not inherit them.

Circularity Check

0 steps flagged

No significant circularity; model outputs and empirical gradients are data-driven and cross-checked against independent diagnostics rather than forced by construction or self-citation.

full rationale

The paper's derivation chain is self-contained and non-circular. SPEI construction evaluates three external RET formulas and five candidate distributions (BATs from Stein 2021; GEV/GenLog/Pe-III/Normal) via independent normality tests (SW/AD/Lilliefors with FDR) on the standardized series; acceptance rates and near-identical spatially-averaged SPEI-3/6/12 curves across RET methods are empirical outcomes, not definitional. The duration-augmented hazard (Eq. 8) extends the BMCD of Doizé et al. (2026) with independent GP priors and is fitted by HMC; recovery/survival fields and the β_total map are direct posterior functionals of that fit, not re-labeled inputs. Predictive superiority is assessed by LOOIC/WAIC and residual Moran’s I against a non-spatial baseline; the south–north persistence gradient is further corroborated by independent Hurst/DFA maps computed on the same SPEI series. No load-bearing uniqueness theorem, self-citation chain, or fitted-parameter-as-prediction appears; parametric form of the hazard is an explicit modeling choice whose qualitative regional contrast survives the paper’s own external checks.

Axiom & Free-Parameter Ledger

6 free parameters · 6 axioms · 1 invented entities

The central persistence claim rests on a binary drought definition, a parametric duration-hazard link, GP smoothness, and standard SPEI construction choices. Free parameters are hierarchical priors and GP hyperparameters fitted by HMC; axioms are mostly domain standards plus the BMCD–renewal equivalence imported from Doizé et al. No new physical entity is postulated—BATs and spatial BMCD are modeling constructs with independent prior literature, so invented_entities is empty of particles/forces but records the paper-specific spatial hazard field as a constructed object.

free parameters (6)
  • GP range parameters ρ_α, ρ_β
    Length scales of baseline and slope GPs; Gamma(2, λ) priors with mean 0.25× domain extent; posterior values drive spatial smoothness of persistence.
  • GP marginal SDs σ_α, σ_β
    Inv-Gamma(2.5, 0.5) priors; control how much local hazard can deviate from global α0, β0 on the logit scale.
  • Fixed effects α0, β0
    Global intercept and log-duration slope with N(0,4) and N(0,1) priors; set the domain-average recovery baseline and duration dependence.
  • BATs parameter vector θ = (α_i, β_i, γ_i) per grid and scale
    Location, scale, and dual tail indices fitted by MLE at each of 730 cells and each accumulation scale before SPEI standardization.
  • Drought threshold SPEI < −1
    Binary state definition used for all persistence modeling; conventional but not estimated from data and load-bearing for spell construction.
  • HSGP basis count m and bounded domain
    Approximation rank for the spatial GPs; affects fidelity of β(s) and α(s) fields but is not reported as a sensitivity study.
axioms (6)
  • domain assumption BMCD exit probabilities are equivalent to an alternating renewal process so that spell durations are iid under the construction (Doizé et al. 2026).
    Imported as the foundation for duration-dependent recovery hazards (§2.5); not re-proved for drought SPEI series.
  • domain assumption Valid SPEI series are those whose standardized values are acceptably standard normal (SW/AD/Lilliefors after FDR control).
    Evaluation criterion for candidate distributions (§2.4); standard in the SPEI literature but still an assumption about what 'good index' means.
  • ad hoc to paper Logit recovery hazard is linear in log(duration+1) with additive spatial random intercept and slope (Eq. 8).
    Core parametric form of the spatial model; not derived from physical first principles.
  • domain assumption Matérn 5/2 GP priors produce appropriate smoothness for environmental hazard fields; separate ranges for α and β.
    §2.5.1; standard spatial-stats choice (Stein 1999) used without alternative kernels.
  • domain assumption First and last drought spells in each record segment are discarded to avoid boundary censoring.
    Follows Doizé et al.; reduces sample size and can bias long-spell representation if many spells are truncated.
  • domain assumption Climatic water balance D = P − RET is the correct input process for multi-scalar drought standardization.
    SPEI definition (Vicente-Serrano et al. 2010); taken as given.
invented entities (1)
  • Spatial total slope field β_total(s) = β0 + β(s) as a continuous persistence regime map no independent evidence
    purpose: Summarize location-specific duration dependence and classify persistent vs memoryless drought regimes across Italy.
    Constructed posterior functional of the hierarchical model, not a new physical quantity; independent_evidence is false because it is defined only inside this fit.

pith-pipeline@v1.1.0-grok45 · 33159 in / 3712 out tokens · 48623 ms · 2026-07-10T20:56:23.371408+00:00 · methodology

0 comments
read the original abstract

Drought monitoring across heterogeneous climate regions requires explicit modeling of drought persistence rather than relying solely on index classification. We develop a Bayesian spatial, duration-augmented framework for drought persistence that captures both temporal memory and spatial dependence. Because persistence estimates depend on the underlying drought index, we also assess the influence of (i) reference evapotranspiration (RET) formulations and (ii) the choice of probability distribution used to standardize the climatic water balance. The framework is applied to a high-resolution (0.25$^\circ \times$ 0.25$^\circ$) dataset over 730 grid cells across Italy. The choice of RET formulation (Penman-Monteith, Hargraves, Thornthwaite) has only a marginal influence on the overall drought signal: spatially averaged SPEI series at the 3-, 6-, and 12-month scales are identical across methods. By contrast, the proposed bulk-and-tails (BATs) distribution substantially improves index estimation, consistently outperforming commonly used alternative models across all RET methods and timescales and achieving normality acceptance rates of 98.4%--99.9% compared to 37%--92% for the generalized extreme value (GEV) distribution and 0%--30% for others. The spatial duration-augmented model further outperforms a nonspatial counterpart, producing spatially continuous, uncertainty-aware recovery and survival fields. Results reveal a strong south-north gradient in temporal memory: southern peninsular and island regions (Sicily, Sardinia) exhibit pronounced persistence and duration-dependent recovery hazards, whereas northern and Alpine regions show near-zero duration dependence. Drought characterization is more sensitive to the distributional and persistence assumptions than to the RET formulation, and duration-insensitive classifications systematically understate extended drought risk in southern Italy.

Figures

Figures reproduced from arXiv: 2607.06805 by Thordis Linda Thorarinsdottir, Touqeer Ahmad.

Figure 1
Figure 1. Figure 1: Study area spatial map. strong spatial variability, reflecting Italy’s complex topography and Mediterranean climatic con￾ditions. A clear north-to-south variability appears in both years. Northern Italy, particularly the Alpine and Po Valley regions, shows lower temperatures, higher precipitation, and higher rela￾tive humidity, whereas southern Italy and the islands are warmer and drier, with higher solar … view at source ↗
Figure 2
Figure 2. Figure 2: Spatial distribution of variables for June 1990 and June 2010: (a) mean maximum temperature, (b) mean solar radiation, (c) mean wind speed, (d) mean relative humidity, (e) precipitation sum, and (f) mean FAO Penman–Monteith RET. 7 [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Spatial distribution of RETj , j ∈ PM, TW, HG at one month scale. Recovery and Survival Probability Curves: For each location s and duration d = 1, . . . , months, the posterior recovery probability for draw k is h (k) (s, d) = logit−1 [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Spatial distribution of correlation among the monthly summed RETj , j ∈ {PM, TW, HG}. of Italy, with the strongest effects in southern regions and along coastal zones. In contrast, RETTW produced lower RET estimates and relatively smoother spatial transitions, especially across the Alpine and semi-arid regions. Figure S.1 shows spatial distribution of the differences of RETTW and RETHG with RETPM. RETTW is… view at source ↗
Figure 5
Figure 5. Figure 5: Shapiro–Wilk normality test acceptance rate (%) for SPEI obtained through different candidate distributions for time scales 1 to 12. Left to right: RETPM, RETTW, and RETHG methods were used in D. The temporal evolution of the spatially averaged SPEI-3, SPEI-6 and SPEI-12 calculated via BATs models ( [PITH_FULL_IMAGE:figures/full_fig_p015_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Spatial comparison of the SW test at each grid cell under RETPM for BATs-SPEI-1 through BATs-SPEI-12. conditions, while negative slope values correspond to precipitation deficit and highlight increasing drought conditions. In a short period (e.g., SPEI-3), positive and statistically significant trends prevail in some areas of southern Italy, i.e., Campania, Calabria, Basilicata, and northern Sicily, indica… view at source ↗
Figure 7
Figure 7. Figure 7: Temporal variation of spatially averaged BATs-SPEI time series over Italy calculated at 3-, 6-, and 12-month accumulation scales based on the RETPM, RETTW, and RETHG. a strong meridional gradient in the long-term memory behavior. The largest H values, typically ranging from 0.73 to 0.81, are observed in the southern peninsular regions and major islands, including extensive areas of Calabria, Sicily, and Sa… view at source ↗
Figure 8
Figure 8. Figure 8: Spatial distribution of trends for BATs-SPEI-3, SPEI-6 and SPEI-12 under RETPM. 38°N 40°N 42°N 44°N 46°N Latitude 8°E 10°E 12°E 14°E 16°E 18°E Longitude Hurst 0.64 0.68 0.72 0.76 38°N 40°N 42°N 44°N 46°N Latitude 8°E 10°E 12°E 14°E 16°E 18°E Longitude α 0.70 0.75 0.80 0.85 0.90 0.95 Spatial distribution of long−range dependence metrics for SPEI−6 across Italy [PITH_FULL_IMAGE:figures/full_fig_p018_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Spatial distribution of long-range dependence and scaling metrics: (left) Hurst exponent; (right) detrended fluctuation analysis exponent for BATs-SPEI-6 under RETPM. high-frequency variability, consistent with fast hydroclimatic fluctuations and limited drought memory. Conversely, DFA exponents exceeding 0.85, found in southern Italy, eastern Sicily, northeastern Apulia, and Sardinia, indicate very strong… view at source ↗
Figure 10
Figure 10. Figure 10: Spatial and temporal patterns of drought persistence across Italy. (a) Spatial distribu￾tion of the total duration-dependence parameter, βtotal(s). (b) Posterior mean drought recovery probability curves. (c) Corresponding drought survival probability curves. 12 months than a geometric distribution would imply. Risk assessments based on memoryless models will therefore systematically underestimate the prob… view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

77 extracted references · 77 canonical work pages

  1. [1]

    Journal of Environmental Sciences , volume=

    Climate change impacts on eutrophication in the Po River (Italy): Temperature-mediated reduction in nitrogen export but no effect on phosphorus , author=. Journal of Environmental Sciences , volume=. 2024 , publisher=

    work page 2024

  2. [2]

    Statistics and Computing , volume=

    Practical Hilbert space approximate Bayesian Gaussian processes for probabilistic programming , author=. Statistics and Computing , volume=. 2023 , publisher=

    work page 2023

  3. [3]

    Journal of the Royal statistical society: series B (Methodological) , volume=

    Controlling the false discovery rate: a practical and powerful approach to multiple testing , author=. Journal of the Royal statistical society: series B (Methodological) , volume=. 1995 , publisher=

    work page 1995

  4. [4]

    The anatomy of drought in Italy: statistical signatures, spatiotemporal persistence, and forecasting potential , journal =

    Francesco Granata and Fabio. The anatomy of drought in Italy: statistical signatures, spatiotemporal persistence, and forecasting potential , journal =. 2026 , issn =. doi:https://doi.org/10.1016/j.jhydrol.2025.134428 , url =

    work page doi:10.1016/j.jhydrol.2025.134428 2026

  5. [5]

    Journal of Machine Learning Research , year =

    Watanabe, Sumio , title =. Journal of Machine Learning Research , year =

    work page

  6. [6]

    Journal of Hydrology , pages=

    Exploring soil moisture dynamics and variability across scales and geological settings using gaussian mixture-long short-term memory networks , author=. Journal of Hydrology , pages=. 2025 , publisher=

    work page 2025

  7. [7]

    Weather and Climate Extremes , pages=

    Identifying Recurring Patterns of Extreme Daily Precipitation Using K-means algorithm: Uncovering Spatial Shift driven by Climate Change over the Italian Peninsula , author=. Weather and Climate Extremes , pages=. 2025 , publisher=

    work page 2025

  8. [8]

    npj Climate and Atmospheric Science , volume=

    Global climatology and trends in convective environments from ERA5 and rawinsonde data , author=. npj Climate and Atmospheric Science , volume=. 2021 , publisher=

    work page 2021

  9. [9]

    Journal of Hydrology , volume=

    Event-based rainfall dynamics and storm frequency in arid and semi-arid climates: An innovative standardized approach for rainstorm analysis , author=. Journal of Hydrology , volume=. 2025 , publisher=

    work page 2025

  10. [10]

    Computers, Environment and Urban Systems , volume=

    Practical limits of Moran’s autocorrelation index for raster class maps , author=. Computers, Environment and Urban Systems , volume=. 2007 , publisher=

    work page 2007

  11. [11]

    Porcu, Emilio and Bevilacqua, Moreno and Schaback, Robert and Oates, Chris J , journal=. The Mat. 2024 , publisher=

    work page 2024

  12. [12]

    EGUsphere , volume=

    Characterizing Soil Moisture Drought Onset and Termination in Europe , author=. EGUsphere , volume=. 2025 , publisher=

    work page 2025

  13. [13]

    2005 , publisher=

    An Introduction to Markov Processes , author=. 2005 , publisher=

    work page 2005

  14. [14]

    Atmospheric Research , volume=

    Drought analysis and water resource availability using standardised precipitation evapotranspiration index , author=. Atmospheric Research , volume=. 2018 , publisher=

    work page 2018

  15. [15]

    Journal of Hydrology: Regional Studies , volume=

    Spatial variability of precipitation extremes over Italy using a fine-resolution gridded product , author=. Journal of Hydrology: Regional Studies , volume=. 2021 , publisher=

    work page 2021

  16. [16]

    International journal of biometeorology , volume=

    Temperature humidity index scenarios in the Mediterranean basin , author=. International journal of biometeorology , volume=. 2013 , publisher=

    work page 2013

  17. [17]

    International Journal of Climatology , volume=

    Climatology, variability, and trends in near-surface wind speeds over the North Atlantic and Europe during 1979--2018 based on ERA5 , author=. International Journal of Climatology , volume=. 2021 , publisher=

    work page 1979

  18. [18]

    Applied Geomatics , volume=

    Recent (1981--2024) climatic trends in Italy from high-resolution reanalysis data and station-based observations , author=. Applied Geomatics , volume=. 2026 , publisher=

    work page 1981

  19. [19]

    Regional Environmental Change , volume=

    Hydrological extremes in the Mediterranean basin: interactions, impacts, and adaptation in the face of climate change , author=. Regional Environmental Change , volume=. 2025 , publisher=

    work page 2025

  20. [20]

    Ecological Indicators , volume=

    Olive grove landscape change: A spatial analysis using multitemporal geospatial datasets , author=. Ecological Indicators , volume=. 2024 , publisher=

    work page 2024

  21. [21]

    Natural Hazards , volume=

    Influence of climate change on irrigation demand: insights from one of the most agricultural regions in Italy (Puglia) , author=. Natural Hazards , volume=. 2025 , publisher=

    work page 2025

  22. [22]

    Italian Journal of Agronomy , volume=

    Climate change impacts on the Alpine, Continental and Mediterranean grassland systems of Italy: A review , author=. Italian Journal of Agronomy , volume=. 2021 , publisher=

    work page 2021

  23. [23]

    Physica A: Statistical Mechanics and its Applications , volume=

    Detecting long-range correlations with detrended fluctuation analysis , author=. Physica A: Statistical Mechanics and its Applications , volume=. 2001 , publisher=

    work page 2001

  24. [24]

    Hydrological Sciences Journal , volume=

    Climate change, the Hurst phenomenon, and hydrological statistics , author=. Hydrological Sciences Journal , volume=. 2003 , publisher=

    work page 2003

  25. [25]

    Stein and Adam H

    Mitchell Krock and Julie Bessac and Michael L. Stein and Adam H. Monahan , keywords =. Nonstationary seasonal model for daily mean temperature distribution bridging bulk and tails , journal =. 2022 , issn =. doi:https://doi.org/10.1016/j.wace.2022.100438 , url =

    work page doi:10.1016/j.wace.2022.100438 2022

  26. [26]

    Mathematical Proceedings of the Cambridge Philosophical Society , volume=

    Limiting Forms of the Frequency Distribution of the Largest or Smallest Member of a Sample , author=. Mathematical Proceedings of the Cambridge Philosophical Society , volume=. 1928 , doi=

    work page 1928

  27. [27]

    , title =

    Gnedenko, Boris V. , title =. Annals of Mathematics , volume =. 1943 , doi =

    work page 1943

  28. [28]

    Journal of Climate , volume=

    A multiscalar drought index sensitive to global warming: the standardized precipitation evapotranspiration index , author=. Journal of Climate , volume=

    work page

  29. [29]

    Statistics and computing , volume=

    Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC , author=. Statistics and computing , volume=. 2017 , publisher=

    work page 2017

  30. [30]

    Chapter 9 - East Atlantic and Scandinavian patterns , editor =

    Armand Hernandez and Laia Comas-Bru , keywords =. Chapter 9 - East Atlantic and Scandinavian patterns , editor =. Atmospheric Oscillations , publisher =. 2025 , isbn =. doi:https://doi.org/10.1016/B978-0-443-15638-0.00009-5 , url =

    work page doi:10.1016/b978-0-443-15638-0.00009-5 2025

  31. [31]

    Global and Planetary Change , volume=

    Climate variability in north-western Italy during the second half of the 20th century , author=. Global and Planetary Change , volume=. 2008 , publisher=

    work page 2008

  32. [32]

    Data in Brief , volume=

    MADIA-Meteorological variables for agriculture: A dataset for the Italian area , author=. Data in Brief , volume=. 2023 , publisher=

    work page 2023

  33. [33]

    International journal of climatology , volume=

    Standardized precipitation evapotranspiration index (SPEI) revisited: parameter fitting, evapotranspiration models, tools, datasets and drought monitoring , author=. International journal of climatology , volume=. 2014 , publisher=

    work page 2014

  34. [34]

    Wiley Interdisciplinary Reviews: Water , volume=

    Hydrological drought explained , author=. Wiley Interdisciplinary Reviews: Water , volume=. 2015 , publisher=

    work page 2015

  35. [35]

    Nature Reviews Earth & Environment , volume=

    A typology of compound weather and climate events , author=. Nature Reviews Earth & Environment , volume=. 2020 , publisher=

    work page 2020

  36. [36]

    Statistics and Computing , volume=

    Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC , author=. Statistics and Computing , volume=. 2017 , publisher=

    work page 2017

  37. [37]

    Hydrological Sciences Journal , volume=

    A dynamic approach for assessing climate change impacts on drought: an analysis in Southern Italy , author=. Hydrological Sciences Journal , volume=. 2023 , publisher=

    work page 2023

  38. [38]

    International Journal of Climatology , volume=

    World drought frequency, duration, and severity for 1951--2010 , author=. International Journal of Climatology , volume=

    work page 1951

  39. [39]

    Weather Climate Extreme 46, 100729 , author=

    The 2022 record-high heat waves over southwestern Europe and their underlying mechanism. Weather Climate Extreme 46, 100729 , author=

    work page 2022

  40. [40]

    Theoretical and Applied Climatology , volume=

    Influence of the North Atlantic Oscillation on winter rainfall in Calabria (southern Italy) , author=. Theoretical and Applied Climatology , volume=. 2013 , publisher=

    work page 2013

  41. [41]

    Quaternary Science Reviews , volume=

    Responses of small mountain glaciers in the Maritime Alps (south-western European Alps) to climatic changes during the Last Glacial Maximum , author=. Quaternary Science Reviews , volume=. 2024 , publisher=

    work page 2024

  42. [42]

    Global and Planetary Change , volume=

    Climate change projections for the Mediterranean region , author=. Global and Planetary Change , volume=. 2008 , publisher=

    work page 2008

  43. [43]

    Regional Environmental Change , volume=

    The relation between climate change in the Mediterranean region and global warming , author=. Regional Environmental Change , volume=. 2018 , publisher=

    work page 2018

  44. [44]

    Earth-Science Reviews , volume=

    Investigating soil moisture--climate interactions in a changing climate: A review , author=. Earth-Science Reviews , volume=. 2010 , publisher=

    work page 2010

  45. [45]

    International Journal of Climatology , volume=

    Widespread Multi-Year Droughts in Italy: Identification and Causes of Development , author=. International Journal of Climatology , volume=. 2025 , publisher=

    work page 2025

  46. [46]

    MADIA - Meteorological variables for agriculture: A dataset for the Italian area , journal =

    Barbara Parisse and Roberta Alilla and Antonio Gerardo Pepe and Flora. MADIA - Meteorological variables for agriculture: A dataset for the Italian area , journal =. 2023 , issn =. doi:https://doi.org/10.1016/j.dib.2022.108843 , url =

    work page doi:10.1016/j.dib.2022.108843 2023

  47. [47]

    International Journal of Climatology , volume=

    Candidate distributions for climatological drought indices (SPI and SPEI) , author=. International Journal of Climatology , volume=. 2015 , publisher=

    work page 2015

  48. [48]

    Environmetrics , volume=

    A parametric model for distributions with flexible behavior in both tails , author=. Environmetrics , volume=. 2021 , publisher=

    work page 2021

  49. [49]

    Journal of Hydrology: Regional Studies , volume=

    Sensitivity of Standardized Precipitation and Evapotranspiration Index (SPEI) to the choice of SPEI probability distribution and evapotranspiration method , author=. Journal of Hydrology: Regional Studies , volume=. 2024 , publisher=

    work page 2024

  50. [50]

    Food and Agriculture Organization of the United Nations, Rome , volume=

    Crop evapotranspiration-Guidelines for computing crop water requirements-FAO Irrigation and drainage paper 56 , author=. Food and Agriculture Organization of the United Nations, Rome , volume=

    work page

  51. [51]

    and Wilk, Martin B

    Shapiro, Samuel S. and Wilk, Martin B. , title =. Biometrika , year =

    work page

  52. [52]

    and Darling, Donald A

    Anderson, Theodore W. and Darling, Donald A. , title =. Annals of Mathematical Statistics , year =

    work page

  53. [53]

    , title =

    Stephens, Michael A. , title =. Journal of the American Statistical Association , year =

    work page

  54. [54]

    , title =

    Lilliefors, Hubert W. , title =. Journal of the American Statistical Association , year =

    work page

  55. [55]

    , title =

    Mann, Henry B. , title =. Econometrica , year =

    work page

  56. [56]

    , title =

    Kendall, Maurice G. , title =. 1975 , edition =

    work page 1975

  57. [57]

    , title =

    Sen, Pranab K. , title =. Journal of the American Statistical Association , year =

    work page

  58. [58]

    and Box, George E

    Ljung, Greta M. and Box, George E. P. , title =. Biometrika , year =

    work page

  59. [59]

    Journal of Hydrology , volume=

    A modified Mann-Kendall trend test for autocorrelated data , author=. Journal of Hydrology , volume=. 1998 , publisher=

    work page 1998

  60. [60]

    goodness of fit

    Asymptotic theory of certain" goodness of fit" criteria based on stochastic processes , author=. The annals of mathematical statistics , pages=. 1952 , publisher=

    work page 1952

  61. [61]

    Science of the Total Environment , volume=

    Assessment of evaporative demand drought index for drought analysis in Peninsular Malaysia , author=. Science of the Total Environment , volume=. 2024 , publisher=

    work page 2024

  62. [62]

    Lilliefors , title =

    Hubert W. Lilliefors , title =. Journal of the American Statistical Association , volume =. 1967 , publisher =. doi:10.1080/01621459.1967.10482916 , URL =

    work page doi:10.1080/01621459.1967.10482916 1967

  63. [63]

    brms: An R Package for Bayesian Multilevel Models Using Stan , journal =

    B. brms: An R Package for Bayesian Multilevel Models Using Stan , journal =. 2017 , doi =

    work page 2017

  64. [64]

    and Lee, Daniel and Goodrich, Ben and Betancourt, Michael and Brubaker, Marcus and Guo, Jiqiang and Li, Peter and Riddell, Allen , title =

    Carpenter, Bob and Gelman, Andrew and Hoffman, Matthew D. and Lee, Daniel and Goodrich, Ben and Betancourt, Michael and Brubaker, Marcus and Guo, Jiqiang and Li, Peter and Riddell, Allen , title =. Journal of Statistical Software , volume =. 2017 , doi =

    work page 2017

  65. [65]

    arXiv preprint arXiv:2605.20399 , year=

    A Duration-Augmented Binary Markov Chain for Rainfall Occurrence with Long Dry Spells , author=. arXiv preprint arXiv:2605.20399 , year=

    work page arXiv

  66. [66]

    Bayesian Analysis , volume =

    Gelman, Andrew , title =. Bayesian Analysis , volume =. 2006 , doi =

    work page 2006

  67. [67]

    Gelman, Andrew and Hill, Jennifer , title =

    work page

  68. [68]

    The Annals of Applied Statistics , volume =

    Gelman, Andrew and Jakulin, Aleks and Pittau, Maria Grazia and Su, Yu-Sung , title =. The Annals of Applied Statistics , volume =. 2008 , doi =

    work page 2008

  69. [69]

    and Stern, Hal S

    Gelman, Andrew and Carlin, John B. and Stern, Hal S. and Dunson, David B. and Vehtari, Aki and Rubin, Donald B. , title =

    work page

  70. [70]

    Distributions of Long Dry Spells in the Iberian Peninsula, Years 1951--1990 , journal =

    Lana, Xavier and Burgue. Distributions of Long Dry Spells in the Iberian Peninsula, Years 1951--1990 , journal =. 2006 , doi =

    work page 1951

  71. [71]

    and Sanchez, E

    Lopez-Franca, N. and Sanchez, E. and Losada, T. and Dominguez, M. and Romera, R. and Gaertner, M. A. , title =. Climate Dynamics , volume =. 2015 , doi =

    work page 2015

  72. [72]

    , title =

    Resnick, Sidney I. , title =

    work page

  73. [73]

    and Willett, John B

    Singer, Judith D. and Willett, John B. , title =

    work page

  74. [74]

    , title =

    Stein, Michael L. , title =

    work page

  75. [75]

    Stern, R. D. and Coe, R. , title =. Journal of the Royal Statistical Society: Series A (General) , volume =. 1984 , doi =

    work page 1984

  76. [76]

    Statistics and Computing , volume =

    Vehtari, Aki and Gelman, Andrew and Gabry, Jonah , title =. Statistics and Computing , volume =. 2017 , doi =

    work page 2017

  77. [77]

    , title =

    Wilks, Daniel S. , title =. Agricultural and Forest Meteorology , volume =. 1999 , doi =

    work page 1999