Climate Change, Climate Variability, & Drought Portfolio

Haiku Central, Maui

2024-06-04

Climate change, climate variability, and drought (CCVD) will exert a growing impact on Hawaii's ecosystems, agriculture and communities in the future. While resource managers are tasked with preparing for this with the best available information, it is hard to know what data, research and recommendations are available. The Pacific Drought Knowledge Exchange (PDKE) program focuses on facilitating knowledge exchange between the research community and resource managers and stakeholders, thereby expanding the utility of climate and drought-related scientific products.��This CCVD portfolio is a comprehensive synthesis of climate and drought information developed specifically for Haiku Central (Central). It is designed to provide relevant climate and drought information needed to inform land management and guide future research and extension. While we try to include a wide range of useful site-specific data products, we also recognize that every site is unique and PDKE is happy to collaborate on producing additional drought products beyond the CCVD portfolio to meet stakeholder needs.� �The PDKE program was piloted in November of 2019 with funding from the Pacific Islands Climate Adaptation Science Center (PICASC). Subsequent PDKE activities and updates to the CCVD portfolio have been funded by PICASC, the East-West Center, and the National Integrated Drought Information System (NIDIS).

2

Longman et al. (2022)

Part 1: Describing the Area

In describing any area of management in Hawaii, it is important to present both traditional and contemporary knowledge. Traditional Hawaiian landscape divisions are well-documented and were established largely following geological features and the natural flow of resources throughout the landscape. ��This portfolio provides a brief description of Central in context of traditional Hawaiian landscapes, as well as contemporary knowledge on elevation and current landcover.

Credit: Matt Foster

3

Hawaiian Land Divisions

There are three types of traditional Hawaiian landscape divisions available as GIS layers and presented here for Central. The land divisions are shown in red and Central is shown in blue.

Mokupuni is the largest land division and refers to entire islands. Central is in the mokupuni of Maui.

Within mokupuni are smaller divisions called moku. Central is in the moku of Hamakualoa.

Within each moku are several ahupuaa which commonly extend from uplands to the sea. Central is situated within 10 ahupuaa including Paʻuwela; ʻŌpana; and Kuiaha.

https://planning.hawaii.gov/gis

4

Elevation

Central is located on the Island of Maui. It covers 4970 acres and a vertical elevation range of 1248 ft. In Hawaii, climate gradients can change significantly over short distances due to changes in elevation, topography, and orientation to the prevailing winds.

Elevation Central

Minimum = 245 ft

Mean = 786 ft

Maximum = 1493 ft

Figure 2. Elevation for the Island of Maui with Central outlined in black.

https://www.pacioos.hawaii.edu/

5

Landcover

The three most common types of landcover in Central are Tree Cover, Grass/Shrub, and Developed. Each landcover type exhibits different climate change impacts and management needs.

Figure 3. Bar graph showing amount and percent of each landcover type within Central.

Figure 4. Landcover mapping for the island of Maui with Central outlined in red. The maps shown in the following slides will be for the Central area only.

https://www.usgs.gov/special-topics/lcmap

6

Water Sources

There are 3 aquifers in Central, their characteristics are listed below.��In general, basal aquifers are more susceptible to saltwater intrusion than high level aquifers.

Figure 5. Department of Health aquifer mapping labeled by DOH Aquifer number and colored by hydrology type. Central outlined in red.

Table 1. Aquifer characteristics for Central.

7

Water Sources

There are Stream, Canal/Ditch, Pipeline and Managed Waterway hydrologic features within Central.��Perennial streams are typically reliable water sources, while intermittent streams stop flowing during dry periods. Well-managed waterways and canal/ditch systems can reduce the impacts of drought and flooding.

Figure 6. Streams and hydrologic features with Central outlined in red.

8

Part 2: Climate Characteristics

Central

In developing this Portfolio, we relied on several gridded climate products available for the State of Hawaii. Annual and monthly estimates of rainfall were obtained from the Hawaii Climate Data Portal (HCDP). Gridded estimates of other climate variables were obtained from the UH Manoa Climate of Hawaii data page. We retrieved all the data points that fell within the boundaries of Central from our 250 meter resolution state-wide maps to support the presented analyses.

9

Rainfall Station Locations

Rainfall data for this portfolio were estimated based on measurements made by hundreds of stations across the state.��The closest station to Central is Haiku, part of the HydroNet-UaNet climate station network.

Figure 6. Rainfall station locations across Maui with Central outlined in blue and three closest stations in orange.

Figure 7. Three closest stations to Central with links to more information on the station network. If there are more than three stations at the site, only three will be listed here.

10

Annual Climate Characteristics

Climatic conditions in Hawaii can vary greatly across the landscape. These maps show the variation in select climate variables across Central and the table below has min. and max. values taken from the maps.

Table 2. Minimum and maximum average annual values for selected climate variables from within Central.

Figure 8. Mean annual climate of Central with area average shown in heading of each plot.

11

Giambelluca et al. (2013;2014)

Average Monthly Rainfall

Average monthly rainfall patterns vary over the course of the year. At Central, the highest monthly rainfall is received in March (10 in.) and the lowest monthly rainfall is received in June (4 in.).

Figure 9. Mean monthly rainfall at Central.

Figure 10. Monthly rainfall maps for the wettest (top) and driest (bottom) months.

12

Giambelluca et al. (2013;2014)

Average Monthly Temperature

Average monthly air temperature patterns vary over the course of the year. At Central there is a 6.6 °F annual variation in temperature, with the warmest month of August (74 °F) and the coolest month in February (68 °F).

Figure 11. Mean monthly air temperature at Central with monthly rainfall in the background.

Figure 12. Monthly air temperature maps for the coldest (top) and hottest (bottom) months.

13

Giambelluca et al. (2013;2014)

Average Monthly Climate

Figure 13. Mean monthly rainfall Central with area average shown in heading of each plot.

Figure 14. Mean monthly temperature at Central with area average shown in heading of each plot.

14

Giambelluca et al. (2013;2014)

Average Seasonal Rainfall

Hawaii has two distinct 6-month seasons of rainfall: hooilo (Wet season: November to April) and kau (Dry season: May to October). Average Wet season monthly rainfall across Central is 8 in and Dry season is 4.6 in. These monthly values are in the 75 and 56 percentiles for rainfall across the whole state, respectively.� �Management plans should anticipate and minimize negative impacts of these seasonal rainfall variations.

Figure 15. Average monthly rainfall maps for the wet (top) and dry (bottom) seasons. Central.

15

Giambelluca et al. (2013)

Average Monthly Climate Table

Central

Table 3. Average monthly climate variables characteristics at Haiku Central. Where, RF is rainfall; Min TA is average minimum air temperature Mean TA is average air temperature; Max TA is average maximum air temperature; RH is relative humidity; CF is cloud frequency; ET is evapotranspiration; SM is soil moisture; S is shortwave downward radiation: ANN, is annual total for rainfall and annual average for all other variables.

16

Giambelluca et al. (2013;2014)

Part 3: Climate Variability

Rainfall and temperature in Hawaii can vary greatly from year-to-year due to natural climatic systems such as the El Niño-Southern Oscillation (ENSO). ENSO is a periodic fluctuation of ocean temperatures in the tropical Pacific, and this has a strong influence on rainfall variability. ENSO consists of five phases, as shown in the graph below.

Figure 16. Timeseries of changes in sea surface temperature (SST) and associated ENSO phase from 1950 - 2022. Central.

https://origin.cpc.ncep.noaa.gov/

17

Seasonal Rainfall and ENSO

In Hawaii, the Warm (El Niño) phase typically brings below average rainfall during the wet season, and above average rainfall in the dry season. This pattern is reversed for the Cool (La Niña) phase.��At Central, the wet season during a Strong El Niño is 16% dryer than the long-term wet season average, and the dry season during a Strong La Niña is 23% dryer than average. These patterns influence drought conditions and wildfire susceptibility, and management activities can benefit from incorporating this ENSO-influenced seasonal rainfall variability.

Figure 17. Barplot of average monthly rainfall grouped by season and ENSO phase. Numbers above the bars are how many seasons from 1950 to 2024 fell within each ENSO phase.

18

Long-Term Trends in Rainfall

Linear trends in annual and seasonal rainfall at Central have been calculated over three different periods since 1920 to show long, mid, and short-term trends. The directions of change for the annual plot are shown below.

Figure 18. Rainfall time series (1920-2024) at Central with linear trends. Trendlines with p-value < 0.05 are statistically significant.

Table 4. Direction of trendline for annual average monthly rainfall over three periods within the record.

19

Frazier et al. (2016); Lucas et al. (2022); See Annex I

Trends in Air Temperature

Trends in monthly air temperature have been calculated over a 33-year record at Central. From 1990 to 2023 average annual air temperature has increased by 1°F.

At this site there is an average range of 11.3°F between the highest and lowest temperatures experienced within a single year. The highest monthly temperature of 81.2°F was recorded in September 2019.

Figure 19. 33-year (1990 to 2023) monthly air temperature time series at Central. The linear trend is determined to be statistically significant when the p-value is less than 0.05.

https://www.hawaii.edu/climate-data-portal/

20

Part 4: Drought and Fire History

Central

Drought is a prominent feature of the climate system in Hawaii and can cause severe impacts across multiple sectors. Droughts in Hawaii often result in reduced crop yields, loss of livestock, drying of streams and reservoirs, depletion of groundwater, and increased wildland fire activity. These impacts can cause substantial economic losses as well as long-term damage to terrestrial and aquatic habitats.

21

Five Types of Drought

There are five major types of drought. During droughts there is sometimes a progression from one type to the next. Depending on local factors, these drought types can also happen simultaneously and at different levels of severity.

22

Frazier et al. (2019)

Identifying Droughts Using the

Standard Precipitation Index

The Standardized Precipitation Index (SPI) is one of the most widely used indices for meteorological drought. SPI compares current rainfall with its multi-year average, so that droughts are defined relative to local average rainfall. This standardized index allows wet and dry climates to be represented on a common scale. Here, 100+ years of monthly rainfall are used to used to calculate SPI-12, which compares how a 12-month period compares with all 12-month periods in the record. SPI-12 is a good measure of sustained droughts that affect hydrological processes at Central.

Figure 20. SPI-12 time series (1920-2024) at Central.

23

Frazier et al. (2016); Lucas et al. (2022)

A 100+ Year History of Drought

Negative SPI values (dry periods) are inverted to show a complete drought timeseries at Central. Dashed lines and corresponding color coding indicates instances of Moderate (SPI > 1), Severe (SPI > 1.5), and Extreme (SPI > 2) drought.

A total of 19 droughts were observed over the entire record with a total of 11 drought events of severe strength or greater. The longest drought lasted for a total of 83 months (see Annex III).

Figure 21. SPI time series (1920-2024) (reversed axis) at Central. Dashed lines show, moderate (yellow), severe (red), and extreme (dark red), drought thresholds.

24

Frazier et al. (2016); Lucas et al. (2022)

Short-term vs Long-term Droughts

Figure 22. SPI-3 time series (reversed axis) at Haiku Central.

Figure 23. SPI-12 time series (reversed axis) at Haiku Central.

SPI-3 provides a comparison of rainfall over a specific 3-month period and reflects short-term conditions. SPI-12 provides 12-month comparisions and reflects long-term conditions. It is important to consider both timescales for planning.

As of March 2024 the most recent drought events are as follows:

SPI-3: Currently in extreme drought since Mar 2023. Current drought intensity is 1.8.

SPI-12: Currently in extreme drought since Jan 2021. Current drought intensity is 2.6.

25

Fire Occurrence in Maui

Ecological drought often drives an increase in wildfire occurrence. In Hawaii, wildfires are most extensive in dry and mesic non-native grass and shrublands. During drought events, wildfire risk in these areas increases rapidly. Currently, agricultural abandonment is resulting in increased grass and shrublands. This combined with recurring incidences of drought is expected to increase the risk of future wildfire in Hawaii.

Figure 24. The map shows wildfires that have occurred on the island of Maui between 1999 and 2022.

26

Trauernicht, 2019; Frazier et al. (2022)

Part 5: Future Climate

Central

Global Climate Models are used to predict future changes in rainfall and temperature, simulating future conditions under different scenarios for how much carbon dioxide we emit into the air. Two common scenarios are RCP 4.5 which assumes we reduce our carbon emissions, and RCP 8.5, which is an increased emissions scenario.

Data downscaling is used make these models useful at the local management level. In Hawaii, two types of downscaled projections are available:

Statistical: Available for Mid & End-of-Century Dynamical: Only available for End-of-Century

Both downscaling projections are presented here. These two projections sometimes agree with each other, and other times they provide conflicting results. When viewing the maps, we can observe where these similarities and differences are, for example which areas show reduced rainfall under both projections.

27

See Annex II

Average Rainfall Change

Mid-Century (2040-2070)

Rainfall for Years 2040-2070

Change in Annual Rainfall

-2 to 2 in/year

-2 to 3% change from present

These Statistical Downscaling maps show the projected change in rainfall under RCP 4.5 and 8.5 conditions. At Central, annual rainfall is projected to decrease by 2 inches (RCP 4.5), or increase by 2 inches (RCP 8.5) by mid-century.

Figure 25. Downscaled future rainfall projections (% change from present) at Central by mid-century (2040-2070) using Statistical Downscaling. RCP 4.5 (left) and RCP 8.5 (right).

28

Timm et al., 2015; See Annex II

Average Rainfall Change

End-of-Century (2100)

Rainfall in the Year 2100

Annual

-3 to 5 in/year

-4 to 7 % change from present

These Dynamical and Statistical Downscaling maps show the projected change in rainfall under RCP 4.5 and 8.5 conditions. At Central, annual rainfall is projected to decrease by 3 inches (RCP 4.5). The models do not agree on the direction of change for RCP 8.5.

Figure 26. Downscaled future rainfall projections (% change from present) at Central by end-of-century (2100), using both Dynamical and Statistical downscaling. RCP 4.5 (top row) and RCP 8.5 (bottom row).

29

Timm et al., 2015; Zhang et al., 2016; See Annex II

Average Air Temperature Change

Mid-Century (2040-2070)

Air Temp. for Years 2040-2070

Change in Air Temperature

1.6 to 3.1°F

2 to 4% change from present

These Statistical Downscaling maps show the projected change in air temperature under RCP 4.5 and 8.5 conditions. At Central, air temperature is projected to increase by 1.6°F (RCP 4.5), or increase by 3.1°F (RCP 8.5) by mid-century.

Figure 27. Downscaled future temperature projections (°F change from present) at Central by mid-century (2040-2070) using Statistical Downscaling. RCP 4.5 (left) and RCP 8.5 (right).

30

Timm 2017; See Annex II

Average Air Temperature Change

End-of-Century (2100)

Air Temp. in the Year 2100

Change in Air Temperature

1.8 to 3.3°F

3 to 5 % change from present

These Dynamical and Statistical Downscaling maps show the projected change in air temperature under RCP 4.5 and 8.5 conditions. At Central, air temperature is projected to increase by 1.8°F (RCP 4.5), or increase by 3.3°F (RCP 8.5) by end-of-century.

Figure 28. Downscaled future rainfall projections (% change from present) at Central by end-of-century (2100), using both Dynamical and Statistical downscaling. RCP 4.5 (top row) and RCP 8.5 (bottom row).

31

Timm 2017; Zhang et al., 2016; See Annex II

Part 6: CCVD Summary

Central

Haiku Central (Central) is located on the island of Maui at mean elevation of 786 ft (245 to 1493 ft). Rainfall varies over the course of the year with a maximum of 10°F occurring in March and a minimum of 4°F occurring in June. On average, wet season months (Nov-Apr) receive 3.4 in more rainfall than dry season months (May-Oct). Seasonal rainfall can vary within the unit as well, with dry season rainfall ranging from 18 to 37°F and wet season rainfall ranging from 34 to 55°F across the 1248 ft elevation gradient. Rainfall can also vary considerably from year-to-year with the driest years occurring during a Strong El Niño event, when on average, 16% less rainfall is received, relative to the long-term average. The average temperature at Central is 71.1°F but temperature ranges from 68°F to 74°F over the course of the year. Drought is a reoccurring feature in the climate system of Central with a total of 19 occurring over the record which is approximately 1.9 per decade. A total of 11 drought events were at severe strength or greater and the longest drought lasted for a total of 83 consecutive months. Future projections of rainfall are uncertain, with end-of-century annual changes ranging from -4 to 7. Future projections of temperature suggest an increase of 1.6°F to 3.1°F by mid century (2040-2070) and an increase of 1.8°F to 3.3°F by the end of the century (2100).

32

External Resources

For more Information

US Drought Monitor

https://droughtmonitor.unl.edu/

NIDIS Current Hawaii Drought Maps

https://www.drought.gov/states/hawaii

State of Hawaii Drought Plan

https://files.hawaii.gov/dlnr/cwrm/planning/HDP2017.pdf

Hawaii Climate Data Portal

https://www.hawaii.edu/climate-data-portal/

ENSO Current Phase and Discussion

https://www.cpc.ncep.noaa.gov/products/analysis_monitoring/enso_advisory/ensodisc.shtml

Pacific Drought Knowledge Exchange

http://www.soest.hawaii.edu/pdke/

Pacific Fire Exchange

https://www.pacificfireexchange.org/

Ahupuaa GIS Layer Storymap

https://storymaps.arcgis.com/stories/1ad9fcf7f2c345a58adef0997fce9b5d

33

Acknowledgements

Ryan Longman, Abby Frazier, Christian Giardina, Derek Ford, Keri Kodama, Alyssa Anderson (PDKE), Mari-Vaughn Johnson, Heather Kerkering, Patrick Grady (PI-CASC), Elliott Parsons (PI-RISCC), Clay Trauernicht, Melissa Kunz, Cherryle Heu (NREM, UH Manoa), Amanda Sheffield, Britt Parker, John Marra (NOAA), Sean Cleveland, Jared McLean (ITS, UH Manoa), Katie Kamelamela (Arizona State University), Thomas Giambelluca (WRRC, UH Manoa), Jim Poterma (SOEST, UH Manoa), Carolyn Auweloa (NRCS), Nicole Galase (Hawaii Cattlemen's Council), Mark Thorne (CTAHR, UH Manoa)

Suggested Citation

Longman, R.J., Ford, D.J., Anderson, A., Frazier, A.G, Giardina, C.P (2023). Climate Change, Climate Variability and Drought Portfolio Haiku Central Pacific Drought knowledge Exchange CCVD Series Version 5.2.

For the most up-to-date version of this portfolio contact Derek Ford: fordd@eastwestcenter.org for more information.

34

Works Cited

35

Annex I: 100+ Year Rainfall

The 100+ year monthly rainfall dataset was drawn from two unique gridded products. We used data from Frazier et al. (2016) for the period 1920-1989 and Lucas et al. (2022) for the period 1990-2024. Given that two unique data sets and methods were used to make these two products we show the 1:1 Statistical relationship between the two products for a 23-year overlap (1990-2012) with the datasets and associated error metrics.

Figure A1: One to one comparison of 23-years (1990-2012) of monthly rainfall from two unique datasets for Central, and associated error metrics; R2, is the coefficient of determination, MBE, is the mean bias error, MAE, mean absolute error.

Frazier et al., 2016; Lucas et al., (2022)

Annex II: Climate Downscaling in Hawaii

Two types of downscaling products were used in this analysis. Here we explain some of the nuances between the two. Dynamical Downscaling (Zhang et al., 2016), feeds GCM output into a regional model that can account for local topographic and atmospheric phenomena at much finer resolutions (e.g. 1 km). End-of-century projections (2100) encompass the period 2080-2099. Statistical Downscaling (Timm et al., 2015, Timm, 2017), develops a relationship between GCM model output and station data for a historical period and then uses this established relationship to make projections for two future scenarios. End-of-century projections (2100) encompass the period 2070-2099 (2100), Mid-century projections encompass the period 2040-2070.

37

Annex III: Drought Events (1920 - 2024)

Table A1. SPI-12 drought characteristics at Haiku Central identified in the SPI-12 timeseries. Duration is the number of months the drought persisted; Average Intensity is the average absolute SPI; Peak Intensity is the highest SPI value calculated during the drought Magnitude is sum of absolute SPI values during the drought.

38

Annex III: Drought Events (1920 - 2024)

Table A1. SPI-12 drought characteristics at Haiku Central identified in the SPI-12 timeseries. Duration is the number of months the drought persisted; Average Intensity is the average absolute SPI; Peak Intensity is the highest SPI value calculated during the drought Magnitude is sum of absolute SPI values during the drought.

39