Tropical cyclones modulate drought characteristics in the Hurricane Region of the Americas

doi:10.21203/rs.3.rs-5276915/v1

Download PDF

Article

Tropical cyclones modulate drought characteristics in the Hurricane Region of the Americas

https://doi.org/10.21203/rs.3.rs-5276915/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Drought and tropical cyclones (TCs) are among the deadliest and costliest natural hazards, and they are expected to intensify in the twenty-first century because of anthropogenic climate change. The Hurricane Region of the Americas (HRA), an area often impacted by TCs and drought of the Americas, hosts some of the most vulnerable countries to these hazards and climate change worldwide. While TCs and drought have been extensively studied separately, there is little research on their interplay in the HRA, especially in areas without quality, long climate data. Here, we analyze the effects of TCs on drought characteristics (e.g., severity and duration) in the HRA between 1985 and 2023 using high-resolution gridded climate data and an array of drought metrics. Our results yield the first-of-its-kind estimate of the interplay between TCs and drought across the entire HRA. We find that, while TCs contribute to 4–15% of annual and seasonal mean precipitation across the region, on average, they ameliorated or terminated drought (e.g., an improvement of at least one drought rank on each metric) at least once in ~ 60% of the HRA in a single month in 1985–2023 (averaged estimates from all drought metrics). We suggest an appropriate analysis of TC-drought interactions should consider several drought metrics, even if sophisticated land-surface models are used.

Earth and environmental sciences/Climate sciences

Earth and environmental sciences/Climate sciences/Hydrology

Tropical cyclones (TCs) are a critical component of the global water and energy balance (1), often associated with human, economic, and environmental losses due to wind damage, storm surge, and heavy rainfall (2, 3, 4). Even though the North Atlantic roughly accounts for the 12% of total annual TC activity worldwide, (5), their impacts often cause significant damage in the Americas (6, 7). While TC-associated precipitation usually results in extensive flooding, it is also beneficial in semi-arid and drought-prone regions (8, 9, 10, 11). TCs contribute up to 30–60% of seasonal and annual mean precipitation in many parts of the world, including northern Australia, northern Mexico, and the Philippines (11, 12, 13), playing a critical role in the annual water budget of these regions. Droughts, on the other hand, are periods of reduced moisture supply (e.g., below-normal precipitation) that unfold more slowly (14), often coupled with increased moisture demand lasting from a few weeks to decades (14, 15, 16, 17). The persistent nature of droughts causes losses in agriculture and threatens food security, energy production, and ecosystems (14, 16), and is even linked to human migration (18). Severe droughts are also associated with an increased wildfire risk and ecosystem disturbances including forest mortality and reduced vegetation growth and uptake of atmospheric CO₂ (19, 20, 21).

While the impacts of drought and TCs have been widely studied separately (2, 3, 10, 14, 16, 17, 22, 23, 24), there is little research on their interactions in the Americas (e.g., 8, 9, 11), especially in regions with sparse climate data, and often most impacted, such as the Caribbean and Mesoamerica (10, 2, 25). Paleoclimate and observational studies link the occurrence of TCs to more intense wildfires (26, 27). TC-induced damage to forests increases fuel mass and alters forest structure, which in turn increases the likelihood of intense fires (27). However, the interaction between TCs and drought is less understood. A land model study in eastern North America suggests that while the contribution of TCs to rainfall is relatively small (ranging from less than 6–15%), they reduce overall drought risk, including drought onset, duration, and geographic extent (9). TC is also associated with transient drier conditions by inhibiting convection in areas distant from the TC’s center due to TC-induced subsidence and moisture divergence (11) Although these conditions typically last a few days, persistent far-away TC trajectories could trigger dry spells (11). That is, a period of days to weeks of below-normal precipitation.

The Hurricane Region of the Americas (HRA), which we define as the area frequently affected by TCs in the Americas (e.g., the Caribbean, Mesoamerica, and parts of North and South America; Fig. 1a), includes some of the most vulnerable countries in the world to climate variability and change (25, 28, 29, 30). The Caribbean and Mesoamerica are among the second most disaster-prone regions in the world, most of which are related to TCs, droughts, and floods (31) TCs and droughts accounted for ~ 40% of weather-related disasters in the Caribbean and Mesoamerica between 2000 and 2022 (31) and are expected to worsen in the twenty-first century due to climate change (29). In Latin America and the Caribbean, drought-related disasters alone affected more than 53 out of ~ 660 million people (32) (approximately 8% of the population) and caused more than $13 billion in losses between 2008 and 2018, making drought the most widespread and costly meteorological hazard in the region (31). Drought alone caused losses of $250 billion between 1980 and 2020 in the US, making it the costliest natural disaster in the country (14, 33). Due to the challenge of climate change in the coming decades, the vulnerability of the HRA to drought and TCs, and our limited understanding of the drought-TC nexus in this region, analyzing the interplay between drought and TCs in the HRA has become critical.

Here, we assess the role of TC-induced precipitation on drought modulation and hydroclimate variability in the HRA. To this end, we have calculated a series of drought metrics before and after removing precipitation from TCs between 1985 and 2023. Our results represent the first analysis of TC-drought interactions in the HRA, integrating drought indices and gridded products. They highlight the importance of TCs in modulating drought onset, duration, and severity across the region.

TC-induced precipitation in the Hurricane Region of the Americas

Between 1985 and 2023, the contribution of TCs to the seasonal mean precipitation in the HRA varies from less than 1–54%, with an area-wide average of ~ 5% (Figs. 1a and 2a). The highest contributions are observed in Baja California, exceeding 40%, and along the west coast of Mexico, with ~ 16% on average. The lowest contributions are observed in north-central Mexico, northern South America, and the continental U.S., with 1–2% of the average seasonal and annual precipitation originating from TCs (Fig. 2a). While the largest contributions are observed in northwestern Mexico, the highest TC-associated seasonal and yearly precipitation rates are on the southern coast of Mexico (Fig. 2a). The highest average per year precipitation rate is over 17 mm/month, while the highest seasonal mean precipitation is ~ 30 mm/month region-wide (Fig. 2b). This finding is consistent with previous studies (e.g., Ref. 11), which indicate that tropical cyclones are an important source of moisture for northwestern Mexico and southwestern U.S.

The highest TC contributions to rainfall in the Caribbean occur in the Lesser Antilles, particularly in Saint Kitts and Nevis and Antigua and Barbuda, with up to 15% and 14% (5% and 4.3% region-wide) on seasonal and annual mean precipitation, respectively (Fig. 2a, b). Average TC contributions in the Greater Antilles are similar to the Lesser Antilles, with ~ 5–4% (7.3–4.3 mm/month) on seasonal and annual precipitation means, respectively (Fig. 2a, b). In Central America, TC contributes up to 11.5% (25.7 mm/month) and 10.4% (15 mm/month) of seasonal and annual precipitation, respectively (Fig. 2). On average, the contribution of TCs to seasonal precipitation is 3.6% (7.8 mm/month), while the yearly mean contribution is 3.2% (4.6 mm/month). There is a notable spatial variability in TC contributions over Mexico, with seasonal contributions up to 54% (48.7 mm/month) and annual contributions up to 49% (28.4 mm/month) in the west-northwest coast of the country (Fig. 2). The southern U.S. receives, on average, up to 11% (24.6 mm/month) from TCs between May and November, with a regional mean of 3.6% (4.4 mm/month). The annual tropical cyclone contribution reaches, on average, 9.7% (14.3 mm/month).

TC contributions to seasonal and annual precipitation are significantly lower in 1985–2005 compared to 2005–2023 (Fig. S1). Seasonal contributions for 2005–2023 were up to 30% higher for 1985–2005. Similarly, the annual contributions were 24% higher in 2005–2023. Notably, the 2005–2023 period includes the most active hurricane seasons in the North Atlantic (2005 and 2020), which may partially explain these results. In contrast, a decrease in TC-associated precipitation occurs in the U.S. southern Central Plains (Fig. S1b, c), ranging from − 6 to − 4% on seasonal and annual means, respectively.

Seasonal and annual TC-contribution estimates from MSWEP are consistent with contributions estimated from CHIRPS in terms of the magnitude and spatial patterns (Fig S2). For example, as in CHIRPS, MSWEP estimates indicate seasonal and annual contributions of up to 51% and 40%, respectively, in Baja California and the west coast of Mexico (Fig S2). However, because MSWEP includes estimates over the ocean, the highest TC contributions to rainfall over the HRA occur in the eastern Pacific Ocean (up to 60%), consistent with the higher TC frequency in this basin compared to the North Atlantic (Fig S2).

TC-contribution estimates from CHIRPS and MSWEP are consistent with weather stations in Mexico, Central America, and southern U.S. in terms of spatial patterns and magnitude (Fig. S3a). For example, Ref. (11) estimated that TCs may contribute up to 40% of the annual precipitation in northwestern Mexico and up to 60% in the Baja California peninsula (Fig. S3a). Further, compared to weather stations across the study area spanning 1997–2013, daily precipitation data from CHIRPS outperform ERA-Interim (ERAI) and Climate Forecast System Reanalysis (CFSR; 34) reanalyses, and Tropical Rainfall Measuring Mission (TRMM; 35), particularly in Northern South America and southern Mexico.

TCs and drought variability in the Hurricane Region of the Americas

Results from one-month scaled Standardized Precipitation minus Evaporation Index (SPEI) before and after subtracting TC-rainfall (SPEI-1_O and SPEI-1_TC, respectively) suggest that between 1985 and 2023, TCs terminated drought (a change in SPEI-1 from ≤ -0.99 to ≥ 0.99 in one month) at least once in ~ 13% (~ one million km²) of the HRA (Table S1). TCs further decreased drought severity (or worsened pluvials) at least once in 67% of the HRA (~ 5.3 million km²), as estimated from a change or at least one drought rank (one unit change) in SPEI-1 (Table S1). The change in drought conditions was statistically significant (p < 0.05 at the 95% level) in approximately 20% of the HRA (Fig. 3a; Table S1). This is consistent with the difference in the kernel density functions and means between SPEI-1_O and SPEI-1_TC, which indicates an increased drought risk after removing TC-precipitation during the months TCs affected the HRA (Fig. 3b,c).

From SPEI-3 and SPEI-6, TCs ended drought in 1.4% and 0.45% of the HRA and decreased at least one drought rank in 43% and 27% (~ 3.4 and 2.1 million km²), respectively (Fig. S4a,b; Table S1). However, these estimates were statistically insignificant. Using SPEI-12, we estimated that TCs ameliorated drought conditions in ~ 17% of the study area, and further ended drought in less than 0.2%, but, again, these changes in drought conditions were not significant (Table S1).

Results from the Standardized Precipitation Index (SPI)-1 indicate that TCs ended drought in 17% of the study area (~ 1.4 million km²) while improving drought conditions in 72%. The change of at least one drought rank in SPI-1 because of TC rainfall was statistically significant (p < 0.05 at the 95% level) in ~ 12% of the HRA (Fig. S4c; Table S1). Similarly, SPI-3 and SPI-6 suggest that TCs ended drought in 3.4 and 1.5% of the HRA, respectively, and increased the indices at least one point (an improvement in drought conditions) in 49 and 33.7% of the area (Table S1). However, these estimates were statistically insignificant. Interestingly, estimates from the self-calibrated Palmer Drought Severity Index (scPDSI; Palmer 1965; Wells et al. 2004) indicate that drought risk decreased in 70% (or ~ 5.5 million km2), where 9% of the area experienced a statistically significant (p > 0.05 at the 95% level) change in scPDSI after the passage of TC during 1985–2023 (Fig. S5). scPDSI estimates further indicate that TCs ended drought at least once in 28% of the HRA (Table S1).

Estimates in TC-drought interactions from the modified Z-score calculated before (Z_O) and after removing TC-precipitation (Z_TC) are consistent with the results obtained using SPEI-1. Both metrics yield similar estimations of TC-induced drought amelioration, both temporal and geographical (Fig. 4a). According to this metric, TCs ended at least one drought in 22% (~ 1.7 million km²) and decreased drought risk (e.g., TCs increased at least 1s of the Z-score) in 76% (~ 5.9 million km²) of the study area (Fig. 4a; Table S1). The change of at least 1s in the Z-score was statistically significant in 41% or about 3.2 million km² of the HRA (Fig. 4.a). While the differences in the distributions of Z_O and Z_TC are similar to SPEI-1_O and SPEI-1_TC, the kernel density functions from Z-scores suggest a greater change in the index after removing the effects of TCs (Fig. 4b). Compared to SPEI-1, for example, subtracting TC-associated rainfall in Z-score shifted the mean from 0.5 to − 0.3s (Fig. 4c).

As climate change is expected to increase the frequency of hydroclimate extremes in the twenty-first century (29, 36), understanding the interplay between TCs and drought is of utmost importance. Here we have comprehensively assessed the influence of TCs on drought characteristics (e.g., drought duration and severity) in the HRA using an array of drought metrics and gridded products. Our results provide the first spatially-consistent analysis of TC’s contributions to precipitation climatology and their effects on drought variability in the region.

Although we used a different period to assess TC-rainfall contributions compared to similar research in the southeastern United States, Mexico, and Mesoamerica, our TC-contribution estimates are consistent with those of previous studies in these regions (3, 9, 11, 13). For example, Ref. (11) estimated TC-contributions to seasonal rainfall of over 40% in arid regions of northern Mexico between 1979 and 2013, while Ref. (3) calculated seasonal contributions of ~ 10% in the US East and Gulf of Mexico coasts between 1948 and 2015. It should be noted that tropical disturbances that did not reach the tropical storm category were not included in the analysis. However, some of these disturbances have been shown to cause significant rainfall events in the HRA. For example, tropical depression One (1990), which formed outside of the typical season (May 24–26), resulted in 150 mm of precipitation in western Cuba (37). This means that our approach could underestimate TC contributions by not including this type of tropical disturbance.

Evaluating the effects of relatively short and episodic events such as TCs on meteorological drought is challenging. Using sophisticated land-surface models is an option to understand the physical processes underpinning TC-drought interactions (9). However, a disadvantage of using these models is that they are not only computationally expensive, but also requires several climate fields and rely on quality observations for accurate simulations. Most areas across the HRA lack long-term quality climate data needed to calibrate most of the current land-surface models, especially the Caribbean and Mesoamerica (10, 11, 25). In this sense, drought indices provide a plausible alternative to evaluate the effects of TCs on drought. Nevertheless, while several indices for assessing drought exist, they are meant for characterizing meteorological drought on time scales from weeks to several years (38, 39, 40). Despite the limitations of current drought metrics in assessing the TC-drought nexus, scalable drought indices such as SPI and SPEI provide an opportunity to better assess the impact of episodic extreme precipitation events on drought if we control for regional and local factors, as we did in this work.

As expected, drought metrics with strong autocorrelations (e.g., the “memory” of past drought conditions) such as SPEI and SPI scaled from three to twelve months, exhibited a lower impact of TC on drought variability. Results from these drought metrics, for example, indicate that less than 1.5% of the study area TCs ended drought at least once between 1985 and 2023 (Table S1). Compared to the Z-scores, SPEI and SPI scaled to one month were more consistent in indicating the role of TCs in modulating drought (Figs. 3, 4). An exception was the estimates using scPDSI (Fig. S5; Table S1), suggesting significant changes in drought severity after the passage of TCs. While scPDSI has a strong autocorrelation from the reliance on the index’s antecedent conditions, how it defines the beginning and end of a drought could partially explain these results. For example, scPDSI backtracks to the previous months to rescale the index when the probability of an established wet or dry interval has reached 100% (38, 41). Although of lesser importance, differences in the available water capacity (AWC) could influence the index’s memory of previous drought conditions. For example, regions with lower AWC exhibit a higher scPDSI variability than those with high AWC (e.g., 10). From Z-scores, SPI-1, and SPEI-1, we estimated that in ~ 22, 17, and 13% of the HRA TCs ended drought at least once between 1985 and 2023 (Table S1).

We used a modified Z-score as a reference drought metric in our analysis because SPI and SPEI (even when scaled to one month) calculate their reference period from the input data. This inherent characteristic may potentially diminish the actual effects of TCs in modulating drought variability. The modified Z-score further avoids the limitations associated with the autocorrelations of SPEI and scPDSI.

While several studies have estimated TC contributions to rainfall on seasonal and annual time scales across the HRA (3, 11, 12, 13, 22), very few have evaluated the role of TCs in modulating drought, most of them in the U.S. and Mexico (e.g., 8, 9, 11). Ref. (10) underscored the importance of TCs in modulating drought severity in the Caribbean Islands. For example, the authors found at least one rank change in scPDSI after the passage of Hurricane David (1979) on Hispaniola Island. Nevertheless, this study did not systematically assess the role of TCs in ending or decreasing drought risk.

Since we did not include the months without the occurrence of landfalling TCs in the HRA in this analysis–even during the hurricane season–our results yield a better estimate of the contribution of TCs in ameliorating or ending drought in the region. This strategy partially offsets the effects of the strong autocorrelation of scPDSI and SPEI-6 in estimating the monthly change in drought due to episodic and intense precipitation only associated with TCs. During the North Atlantic and eastern Pacific hurricane seasons, the HRA is affected by other tropical disturbances (e.g., easterly waves) that cause extreme precipitation events and contribute up to 60% to the seasonal and annual precipitation cycles (e.g., Ref. (42). Assessing the change in drought from a seasonal time scale that includes months without TCs could mask their actual effect on drought. For example, if a particular TC contributed to 80% of the August rainfall in the Caribbean, and easterly waves contributed to 60% of the season. Analyzing the change in drought (e.g., a change in SPEI-1 or Z-score) arising from that particular TC over the entire season, its contribution to change in drought conditions would be lower than isolating and analyzing August alone.

We estimated that hurricane Gustav (2008) and tropical storm Fay (2008), for example, ended a drought on western Hispaniola Island and Cuba within a single month, with changes in SPEI-1 and scPDSI of up to 4.6 and 4.2 units (i.e., from mild drought conditions to moderately wet), respectively (not shown). Similarly, hurricane Florence (2018) ended a mild drought in South Carolina (not shown), with SPEI-1 changing from − 0.27 to 1.45 in a single month. While TCs last a few days, they often bring substantial amounts of precipitation (e.g., 43) that could potentially end a drought (10).

Additionally, the potential effects of TCs in ameliorating or ending other types of droughts, such agricultural droughts, need to be further explored to determine the overall impacts of TCs-related precipitation on these systems. This future assessment should include the impact of TCs-related precipitation on the predictive skill of seasonal precipitation forecasts widely used in the region (44).

Even though we acknowledge the limitations of using drought indices to assess episodic hydroclimate extremes such as TCs, our approach to evaluating the role of TCs in modulating drought variability in the HRA yields results consistent with similar studies using sophisticated land-surface models (e.g., 9) without the computational requirements of them. Nevertheless, our results underscore the critical need to develop metrics and tools capable of capturing the interplay of TC and drought, especially in regions vulnerable to hydroclimate extremes like the Caribbean and Mesoamerica.

Gridded climate data

Table S2 presents a comprehensive list of the gridded datasets used in this work. We used the second version of the Climate Hazards Group Infrared Precipitation with Station Data (CHIRPSv2.0; 45) daily precipitation dataset. This gridded product spans from 1981 to the present, with a native horizontal resolution of 0.05° (~ 6 km). It integrates satellite imagery data with in-situ weather station data covering 50°S to 50°N. CHIRPS is updated monthly for drought monitoring purposes (45). The number of stations used by CHIRPS to calibrate satellite precipitation estimates in our study area is currently ~ 1000, although this number was higher about a decade ago. While there are other gridded precipitation products available, we selected CHIRPS over these data sets for three reasons: (a) CHIRPS provide daily precipitation data at high spatial resolution, (b) CHIRPS is updated regularly, which is critical to assess TCs and drought conditions, (c) compared to weather stations, CHIRPS outperform other high-resolution gridded precipitation products in the Caribbean (46) and Mesoamerica (25, 30).

We also use daily precipitation data from the Multi-Source Weighted-Ensemble Precipitation version 2 (MSWEPv2; 47) as complement to our analysis with CHIRPS. MSWEPv2 merges satellite, ERA5 reanalysis, and ground station data to obtain a global precipitation product with temporal resolutions ranging from 3-hourly to monthly time scales, and a native horizontal resolution of 0.1° (~ 11 km) (47). While this product spans from 1979 to present, we use the 1985–2023 period.

We used monthly potential evapotranspiration (PET) data from the Climatic Research Unit (CRUv4.08). CRU is a widely used gridded product with a native resolution of 0.5°, covering the global land surface except Antarctica (48). The current CRU version spans from January 1901 to December 2023 and is updated annually (48). It is derived by interpolating monthly climate anomalies from extensive networks of weather station observations. CRU uses the UN Food and Agriculture Organization (FAO) reference evapotranspiration (49) to calculate PET, a variant of the Penman-Monteith Eq. (50, 51) that requires fewer climate fields for its computation (49). The concept behind this approach stands from modeling an idealized grass surface with 0.12-m height, a constant water supply, soil resistance of 70 s m^− 1, and a surface albedo of 0.23 (49):

$$\:PET=\:\frac{0.408\varDelta\:\left(Rn-G\right)+\:\gamma\:\frac{900}{T+273.16}{U}_{2}({e}_{s}-{e}_{a})}{\varDelta\:+\:\gamma\:(1+0.34{U}_{2})},$$

where PET is the crop reference evapotranspiration (mm day^− 1), Rn is the net radiation (MJ m^− 2 day^− 1), G is the soil heat flux density (MJ m^− 2 day^− 1), T is the average temperature at 2-m height (°C), U₂ is the wind speed at 2-m height (m s^− 1; estimated from the wind speed measured at 10-m height) e_s − e_a is the vapor pressure deficit for measurement at 2-m height (kPa), Δ is the slope of the vapor pressure curve (kPa °C^− 1), γ is the psychrometric constant (kPa °C^− 1), 900 is the coefficient for the reference crop (kJ^− 1 kg K day^− 1), and 0.34 is the wind coefficient for the reference crop (s m^− 1) (49). We bilinearly interpolated the original PET data from CRU to CHIRPS’ spatial resolution of ~ 6 km.

The gridded available water capacity (AWC) data needed to calculate the scPDSI were obtained from the Global Gridded Surfaces of Selected Soil Characteristics of the International Geosphere-Biosphere Program Data and Information Services (IGBP-DIS) through the Oak Ridge National Laboratory (ORNL; http://daac.ornl.gov/). This database uses a statistical bootstrapping method to generate these surfaces from the FAO-United Nations Educational, Scientific and Cultural Organization (UNESCO) Digital Soil Map of the World (Global Soil Data Task Group 2000). While the native resolution of this product is 5 arcmin, we have pre-processed using a bilinear interpolation routine to interpolate the AWC data to a spatial resolution of ~ 6 km.

HURDAT2

We used the second version of the Hurricane Database (HURDAT2; 52) to determine the tracks and intensities of TCs in the HRA between January 1985 and December 2023. HURDAT2 is frequently updated and covers the North Atlantic and eastern Pacific basins from 1851 to the present (52). While the accuracy of the pre-satellite era (i.e., prior 1980) dataset is limited (3, 53), our study period spans 1985–2023.

Methods and procedures

Study area

We define the Hurricane Region of the Americas (HRA) as the area between 5–40°N and 30–130°W, including the Caribbean, Mesoamerica, Mexico, the southern U.S., and the northern part of South America, spanning an area of about 7.8 million km² (as estimated from the average number of CHIRPS and MSWEP grid cells affected by TC at least once between 1985 and 2023) (Fig. 1a). This area is affected by TCs that form in the North Atlantic and eastern Pacific basins between May and November each year (Fig. 1a).

Estimating TC contributions to precipitation

In Fig. S6, we summarize the procedure we followed for estimating TC contributions to precipitation in the HRA. We estimated the contributions of TC to precipitation in the HRA using daily precipitation data from CHIRPSv2 (45) and MSWEPv2 (47) between 1985 and 2023. We divided the analysis into 20-year periods, 1985–2005 and 2005–2023, to analyze changes and trends in TC contributions to precipitation. While the use of weather station data is often preferred over remotely sensed rainfall estimates (e.g., 13), the lack of long-term and high-quality data prevented us from relying solely on weather stations for this analysis over the entire study area. To assess the performance of CHIRPSv2 and MSWEPv2 in estimating TC-associated precipitation, we compared the TC contributions from these products with quality-assured weather stations from Mexico and Central America for 1997–2013 (e.g., 11). We further compared CHIRPSv2 with MSWEPv2 since both use remotely sensed estimates of precipitation and ground stations to calibrate these estimates (45, 47).

We included only tropical storms and hurricanes as TCs (i.e., maximum sustained winds of at least 39 mph). The rationale for not including tropical depressions (i.e., unnamed TCs) is that they are less likely to have a broad rainfall field based on TC-associated circulation (13). Based on this selection criterion, we did not include subtropical and extratropical storms in this work, which partially avoid collecting extratropical precipitation (3). To determine the tracks of the selected TCs, we used the HURDAT2 dataset for the North Atlantic and eastern Pacific basins. The TCs selected were those whose center was at most 500 km from land. This radial distance has been shown to be appropriate for analyzing TC contributions to local precipitation over the Americas, as topography also plays an important role in producing rainfall, since it is the average radius of TC-associated circulation and rainbands (13). While previous studies have used time-evolving TC radii (54), selecting TCs as tropical storms and hurricanes rather than tropical depressions partially reduces the changes of integrating non-tropical precipitation into our estimates. However, we acknowledge that our approach may under- or overestimate TC-associated precipitation, especially over the US. In addition, since the HURDAT2 data is available every 6 h, we averaged the location of the storm center each day with more than one time step. We used the Python package Tropycalv1.3 to extract and analyze HURDAT2 data.

Finally, we evaluated the contribution of TC to precipitation as the fraction of TC-derived precipitation to total precipitation from monthly, seasonal, and annual contributions. While a monthly CHIRPS dataset is available, we resampled the daily CHIRPS dataset to obtain monthly totals of precipitation for consistency.

Assessing drought

For assessing meteorological drought, we used (a) the Standardized Precipitation Index (SPI; 39), (b) the Standardized Precipitation minus Evaporation Index (SPEI; 40), and (c) the self-calibrated Palmer Drought Severity Index (scPDSI; 38, 41). SPEI and scPDSI employ precipitation and PET to calculate the water balance, thereby enabling the subsequent determination of drought conditions. In contrast, SPI employs precipitation as the sole variable. SPI and SPEI are multiscalar indices, which make them useful for assessing drought at different timescales (e.g., from monthly to annual time scales) (40). In contrast, scPDSI, an improved version of the original PDSI (38), exhibits a strong autocorrelation due to its incorporation of the effect of soil's AWC (38). While these metrics have their respective advantages and limitations, we employed them to obtain more reliable and consistent results. A more comprehensive description of each metric can be found in Ref. (40) for SPEI; Ref. (39) for SPI; and Ref. (41) for scPDSI.

We calculated SPEI and SPI at one-, three-, six-, and 12-month scales to assess the effects of TCs on short-term (one or three month long) to longer term droughts (at least 12 months). We used scPDSI due to its unique ability to account for the influence of soil characteristics on drought variability, which is essential for making comparisons with other studies on drought in the region (e.g., 10, 25). A drought was defined as occurring when the three metrics fell below − 0.99, which is considered to be a moderate drought. All the indices were calculated twice, using CHIRPS before (CHIRPS_O) and after the removal of TC-rainfall (CHIRPS_TC). Because CHIRPS has a resolution of ~ 6 km, we bilinearly interpolated CRU PET data to fit CHIRPS. As with CHIRPS, subscripts _O and _TC are used to represent the index calculated before (CHIRPS_O) and after (CHIRPS_TC) removing TC-associated rainfall.

Assessing TC-drought interactions

We focused our analysis using SPEI and SPI scaled on a monthly timescale, as these metrics exhibit lower autocorrelation. However, to evaluate the longer-term effects of TCs on drought variability, we employed SPEI scaled at three-, six-, and twelve-month intervals and scPDSI.

As SPI, SPEI, and scPDSI derive their calibration period (e.g., the reference period for establishing normal conditions) from the input data, the impact of TC-rainfall on drought is reduced. This means that, while the removal of TC rainfall undoubtedly results in a decrease in total annual precipitation in the HRA, such a reduction is incorporated into the calibration of the drought indices, thereby overshadowing the observed effect. To circumvent this limitation, we calculated a modified Z-score after removing TC rainfall from CHIRPS (i.e., CHIRPS_TC) as:

$$\:{Z}_{TC}=\frac{{CHIRPS}_{TC}-{\mu\:}_{O}}{{\sigma\:}_{O}}$$

where Z_TC is the modified Z-score calculated using CHIRPS after removing TC-rainfall (CHIRPS_TC). The terms 𝜇_o and s_o are the mean and standard deviation of CHIRPS without removing TC rainfall (CHIRPS_O), respectively. In this manner, we preserve the initial variability of

CHIRPS_TC while underscoring the actual effects of TCs in drought variability within the HRA. With all the drought metrics we used (including the modified Z-score), we exclusively evaluated the grid cells affected by TCs between 1985 and 2023. To isolate the grid cells that were directly affected by TCs at a given month during this period, we masked those that were not affected. We then estimated the differences in drought conditions between the areas affected by TCs before and after removing TC-associated rainfall.

Data availability

All the resulting data from this work are publicly available at [https://zenodo.org/uploads/13909808]. The coding used to extract, process, and analyze the data is available at [https://github.com/dimherrera/TC_rainfall_contribution/tree/main]. Additional data and coding related to this work can be requested from the authors.

Author Contribution

D.A.H. conceived and designed the study and prepared the manuscript. C.D. analyzed station data from Mexico and Central America, and A.C. obtained and analyzed MSWEP data. All authors discussed the results and edited the manuscript.

Acknowledgement

This project is partially funded by a start-up funding provided by the University of Tennessee to D.A.H. H.H. wish to acknowledge partial funding for this study through the following Vicerrectoría de Investigación, Universidad de Costa Rica grants: B9454 (supported by Fondo de Grupos), C2103, C4468, C3991 (UCREA), A4906 (PESCTMA) and B0-810. H.H. was partially supported by a grant awarded by the International Development Research Centre (IDRC), Ottawa, Canada, and the Central American University Council (CSUCA-SICA) to the Red Centroamericana de Ciencias sobre Cambio Climático (RC4) project (CR-66, C4468, SIA 0054-2, the opinions expressed here do not necessarily represent those of IDRC, CSUCA, or the Board of Governors). B.I.C. is supported by the NASA Modeling, Analysis, and Prediction program.

Trenberth, K.E. and Fasullo, J., 2007. Water and energy budgets of hurricanes and implications for climate change. Journal of Geophysical Research: Atmospheres, 112(D23).
Keellings, D. and Hernández Ayala, J.J., 2019. Extreme rainfall associated with Hurricane Maria over Puerto Rico and its connections to climate variability and change. Geophysical Research Letters, 46(5), pp.2964–2973.
Bregy, J.C., Maxwell, J.T., Robeson, S.M., Ortegren, J.T., Soulé, P.T. and Knapp, P.A., 2020. Spatiotemporal variability of tropical cyclone precipitation using a high-resolution, gridded (0.25°× 0.25°) dataset for the eastern United States, 1948–2015. Journal of Climate, 33(5), pp.1803–1819.
Ibanez, T., Platt, W.J., Bellingham, P.J., Vieilledent, G., Franklin, J., Martin, P.H., Menkes, C., Pérez-Salicrup, D.R., Russell-Smith, J. and Keppel, G., 2022. Altered cyclone–fire interactions are changing ecosystems. Trends in Plant Science, 27(12), pp.1218–1230.
Emanuel, K., 2021. Atlantic tropical cyclones downscaled from climate reanalyses show increasing activity over past 150 years. Nature communications, 12(1), p.7027.
Dominguez C., Jaramillo A., Cuéllar P. 2021. Are the socioeconomic impacts associated with tropical cyclones in Mexico exacerbated by local vulnerability and ENSO conditions? Int J Climatology. 41 (Suppl. 1): E3307–E3324. https://doi.org/10.1002/joc.6927
Vecchi, G.A., Landsea, C., Zhang, W., Villarini, G. and Knutson, T., 2021. Changes in Atlantic major hurricane frequency since the late-19th century. Nature communications, 12(1), p.4054.
Maxwell, J.T., Ortegren, J.T., Knapp, P.A. and Soulé, P.T., 2013. Tropical cyclones and drought amelioration in the Gulf and southeastern coastal United States. Journal of Climate, 26(21), pp.8440–8452.
Kam, J., Sheffield, J., Yuan, X. and Wood, E.F., 2013. The influence of Atlantic tropical cyclones on drought over the eastern United States (1980–2007). Journal of Climate, 26(10), pp.3067–3086.
Herrera, D. and Ault, T., 2017. Insights from a new high-resolution drought atlas for the Caribbean spanning 1950–2016. Journal of Climate, 30(19), pp.7801–7825.
Dominguez, C. and Magaña, V., 2018. The role of tropical cyclones in precipitation over the tropical and subtropical North America. Frontiers in Earth Science, 6, p.312622.
Breña-Naranjo, J.A., Pedrozo-Acuña, A., Pozos-Estrada, O., Jiménez-López, S.A. and López-López, M.R., 2015. The contribution of tropical cyclones to rainfall in Mexico. Physics and Chemistry of the Earth, Parts A/B/C, 83, pp.111–122.
Khouakhi, A., Villarini, G. and Vecchi, G.A., 2017. Contribution of tropical cyclones to rainfall at the global scale. Journal of Climate, 30(1), pp.359–372.Khouakhi, A., Villarini, G. and Vecchi, G.A., 2017. Contribution of tropical cyclones to rainfall at the global scale. Journal of Climate, 30(1), pp.359–372.
Ault, T.R., 2020. On the essentials of drought in a changing climate. Science, 368(6488), pp.256–260.
Redmond, K.T., 2002. The depiction of drought: A commentary. Bulletin of the American Meteorological Society, 83(8), pp.1143–1147.
Dai, A., 2011. Characteristics and trends in various forms of the Palmer Drought Severity Index during 1900–2008. Journal of Geophysical Research: Atmospheres, 116(D12).
Cook, B.I., Smerdon, J.E., Cook, E.R., Williams, A.P., Anchukaitis, K.J., Mankin, J.S., Allen, K., Andreu-Hayles, L., Ault, T.R., Belmecheri, S. and Coats, S., 2022. Megadroughts in the Common Era and the Anthropocene. Nature Reviews Earth & Environment, 3(11), pp.741–757.
Murray-Tortarolo, G.N., Salgado, M.M. Drought as a driver of Mexico-US migration. Climatic Change 164, 48 (2021). https://doi.org/10.1007/s10584-021-03030-2
Westerling, A. L., Hidalgo, H. G., Cayan, D. R., & Swetnam, T. W., 2006. Warming and earlier spring increase western US forest wildfire activity. science, 313(5789), 940–943. DOI: 10.1126/science.1128.
Humphrey, V., Zscheischler, J., Ciais, P., Gudmundsson, L., Sitch, S. and Seneviratne, S.I., 2018. Sensitivity of atmospheric CO2 growth rate to observed changes in terrestrial water storage. Nature, 560(7720), pp.628–631.
Liu, L., Gudmundsson, L., Hauser, M., Qin, D., Li, S. and Seneviratne, S.I., 2020. Soil moisture dominates dryness stress on ecosystem production globally. Nature communications, 11(1), p.4892.
Knight, D.B. and Davis, R.E., 2007. Climatology of tropical cyclone rainfall in the southeastern United States. Physical Geography, 28(2), pp.126–147.
Knight, D.B. and Davis, R.E., 2009. Contribution of tropical cyclones to extreme rainfall events in the southeastern United States. Journal of Geophysical Research: Atmospheres, 114(D23).
Herrera, D.A., Ault, T.R., Carrillo, C.M., Fasullo, J.T., Li, X., Evans, C.P., Alessi, M.J. and Mahowald, N.M., 2020. Dynamical characteristics of drought in the Caribbean from observations and simulations. Journal of Climate, 33(24), pp.10773–10797.
Anderson, T.G., McKinnon, K.A., Pons, D. and Anchukaitis, K.J., 2023. How exceptional was the 2015–2019 central American drought? Geophysical Research Letters, 50(21), p.e2023GL105391.
LeBlanc, A.R., Kennedy, L.M., Liu, K.B. and Lane, C.S., 2017. Linking hurricane landfalls, precipitation variability, fires, and vegetation response over the past millennium from analysis of coastal lagoon sediments, southwestern Dominican Republic. Journal of Paleolimnology, 58, pp.135–150.
Swann, D.E., Bellingham, P.J. and Martin, P.H., 2024. Cyclone–Fire Interactions Enhance Fire Extent and Severity in a Tropical Montane Pine Forest. Ecosystems, pp.1–18.
Herrera, D.A., Ault, T.R., Fasullo, J.T., Coats, S.J., Carrillo, C.M., Cook, B.I. and Williams, A.P., 2018. Exacerbation of the 2013–2016 pan-Caribbean drought by anthropogenic warming. Geophysical research letters, 45(19), pp.10–619.
IPCC, 2021: Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change[Masson-Delmotte, V., P. Zhai, A. Pirani, S.L. Connors, C. Péan, S. Berger, N. Caud, Y. Chen, L. Goldfarb, M.I. Gomis, M. Huang, K. Leitzell, E. Lonnoy, J.B.R. Matthews, T.K. Maycock, T. Waterfield, O. Yelekçi, R. Yu, and B. Zhou (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, In press, doi:10.1017/9781009157896.
Alfaro-Córdoba, M., Mora‐Sandí, N.P., Hidalgo, H.G. and Alfaro, E.J., 2024. Central American climate extreme trends: A statistical analysis of CLIMDEX indices. International Journal of Climatology, 44(11), pp.4115–4126.
United Nations Office for the Coordination of Humanitarian Affairs (OCHA) & the United Nations Office for Disaster Risk Reduction (UNDRR), 2022. Overview of Disasters in Latin America and the Caribbean 2000–2022. Retrieved from https://www.undrr.org/media/89900/download?startDownload=20240806.
Worldometers.info. 2024.
Smith, A.B. and Matthews, J.L., 2015. Quantifying uncertainty and variable sensitivity within the US billion-dollar weather and climate disaster cost estimates. Natural Hazards, 77, pp.1829–1851.
Saha, S., Moorthi, S., Wu, X., Wang, J., Nadiga, S., Tripp, P., Behringer, D., Hou, Y.T., Chuang, H.Y., Iredell, M. and Ek, M., 2014. The NCEP climate forecast system version 2. Journal of climate, 27(6), pp.2185–2208.
Simpson, J., Adler, R.F. and North, G.R., 1988. A proposed tropical rainfall measuring mission (TRMM) satellite. Bulletin of the American meteorological Society, 69(3), pp.278–295.
Herrera, D.A., Cook, B.I., Fasullo, J., Anchukaitis, K.J., Alessi, M., Martinez, C.J., Evans, C.P., Li, X., Ellis, K.N., Mendez, R. and Ault, T., 2023. Observed changes in hydroclimate attributed to human forcing. PLOS Climate, 2(11), p.e0000303.
Avila, L.A., 1991. Atlantic tropical systems of 1990. Mon. Wea. Rev, 119, pp.2027–2033.
Palmer, W.C., 1965. Meteorological drought (Vol. 30). US Department of Commerce, Weather Bureau.
McKee, T.B., Doesken, N.J. and Kleist, J., 1993, January. The relationship of drought frequency and duration to time scales. In Proceedings of the 8th Conference on Applied Climatology (Vol. 17, No. 22, pp. 179–183).
Vicente-Serrano, S.M., Beguería, S. and López-Moreno, J.I., 2010. A multiscalar drought index sensitive to global warming: the standardized precipitation evapotranspiration index. Journal of climate, 23(7), pp.1696–1718.
Wells, N., Goddard, S. and Hayes, M.J., 2004. A self-calibrating Palmer drought severity index. Journal of climate, 17(12), pp.2335–2351.
Dominguez, C., Done, J.M. and Bruyère, C.L., 2020. Easterly wave contributions to seasonal rainfall over the tropical Americas in observations and a regional climate model. Climate Dynamics, 54(1), pp.191–209.
Sánchez-Murillo, R., Durán-Quesada, A.M., Esquivel-Hernández, G., Rojas-Cantillano, D., Birkel, C., Welsh, K., Sánchez-Llull, M., Alonso-Hernández, C.M., Tetzlaff, D., Soulsby, C. and Boll, J., 2019. Deciphering key processes controlling rainfall isotopic variability during extreme tropical cyclones. Nature communications, 10(1), p.4321.
Acharya, N., Dinku, T. and Hall, K.J.C., 2022. A next generation (NextGen) approach to improve the seasonal prediction system in East Africa. Authorea Preprints. DOI: 10.1002/essoar.10504989.1
Funk, C., Peterson, P., Landsfeld, M., Pedreros, D., Verdin, J., Shukla, S., Husak, G., Rowland, J., Harrison, L., Hoell, A. and Michaelsen, J., 2015. The climate hazards infrared precipitation with stations—a new environmental record for monitoring extremes. Scientific data, 2(1), pp.1–21.
Centella-Artola, A., Bezanilla-Morlot, A., Taylor, M.A., Herrera, D.A., Martinez-Castro, D., Gouirand, I., Sierra-Lorenzo, M., Vichot-Llano, A., Stephenson, T., Fonseca, C. and Campbell, J., 2020. Evaluation of sixteen gridded precipitation datasets over the Caribbean region using gauge observations. Atmosphere, 11(12), p.1334.
Beck, H.E., Wood, E.F., Pan, M., Fisher, C.K., Miralles, D.G., Van Dijk, A.I., McVicar, T.R. and Adler, R.F., 2019. MSWEP V2 global 3-hourly 0.1 precipitation: methodology and quantitative assessment. Bulletin of the American Meteorological Society, 100(3), pp.473–500.
Harris, I., Osborn, T.J., Jones, P. and Lister, D., 2020. Version 4 of the CRU TS monthly high-resolution gridded multivariate climate dataset. Scientific data, 7(1), p.109.
Allen, R.G., 1998. Crop evapotranspiration. FAO irrigation and drainage paper, 56, pp.60–64.
Penman, H.L., 1948. Natural evaporation from open water, bare soil and grass. Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences, 193(1032), pp.120–145.
Monteith, J.L., 1965. Evaporation and environment. In Symposia of the society for experimental biology (Vol. 19, pp. 205–234). Cambridge University Press (CUP) Cambridge.
Landsea, C.W. and Franklin, J.L., 2013. Atlantic hurricane database uncertainty and presentation of a new database format. Monthly Weather Review, 141(10), pp.3576–3592.
Vecchi, G.A. and Knutson, T.R., 2011. Estimating annual numbers of Atlantic hurricanes missing from the HURDAT database (1878–1965) using ship track density. Journal of Climate, 24(6), pp.1736–1746.
Stansfield, A.M., Reed, K.A., Zarzycki, C.M., Ullrich, P.A. and Chavas, D.R., 2020. Assessing tropical cyclones’ contribution to precipitation over the eastern united states and sensitivity to the variable-resolution domain extent. Journal of Hydrometeorology, 21(7), pp.1425–1445.

No competing interests reported.

HerreraetalTCDroughtHRASupMaterials.docx

Download PDF

Reviewers agreed at journal
12 Nov, 2024
Reviewers agreed at journal
11 Nov, 2024
Reviewers invited by journal
25 Oct, 2024
Editor assigned by journal
22 Oct, 2024
Submission checks completed at journal
22 Oct, 2024
First submitted to journal
16 Oct, 2024

You are reading this latest preprint version

Tropical cyclones modulate drought characteristics in the Hurricane Region of the Americas

Status:

Version 1

Abstract

Figures

Introduction

Results

Discussion

Data and Methods

HURDAT2

Methods and procedures

Data availability

Declarations

Author Contribution

Acknowledgement

References

Additional Declarations

Supplementary Files

Status:

Version 1