Welcome to the grass atlas of Hawaiʻi, this site presents data on the geography of all grasses found in Hawaiʻi from herbarium specimens, iNaturalist, vegetation plots, conservation organizations, and roadside surveys by the author. These maps are designed to be used to help understand the elevation ranges at which each species occurs and the approximate distribution of each species across the landscape. If you would like to help contribute to this project, please upload photos of grasses to inaturalist.org and I will both confirm their identity and then add them to this dataset. If you have any questions, comments, want to use any of this information for management desicions, or find errors on this page please email me
DISCLAIMER: The maps presented here DO NOT represent the entire geographic range of each species in Hawaiʻi. For many species, the range shown here is only a small portion of the area it is truly found. As there has never been a rigorous mapping effort for the vegetation of Hawaiʻi, the data presented in this atlas derive from many separate sources which are strongly spatially biased around roads, cities, trails, and National Parks. The models are limited by the input data and will sometimes make erronous predictions, especially predicting areas as unsuitable erronously for invasive species which are still spreading. Although steps have been taken to remove obviously erroneous occurrences, some have likely persisted. No effort was made to remove occurrence data for populations which may have died off. The temporal range of this data is 1900 - 2023. While every effort has been made to ensure the quality and accuracy of these maps, they are not guaranteed for any purpose.
Data sources and methods (show)
Data sources and methods (hide)
Species distribution modelsSpecies distribution modeling was used for some species. This must be thought of as an correlative technique, they largely "fill in gaps" between known occurence points. Therefore, there will be areas which the models predict as unsuitable, but only due to lack of data; once populations from these areas are added to the model, the suitablity will change. This is especially problematic for naturalized species as many are still actively expanding across the landscape. Low modeled suitability values for invasive species do not necessarily mean the area is not suitable. These models are also limited by climate data, as Hawaiʻi's climate has tremendous gradients between windward and leeward areas and there are few weather stations on some of these gradients sometimes leading to unususal or incorect model predictions.
Models were created for species with enough data points to build a reliable model (approximately 40 points). These models were trained using climate data from Hawaiʻi and use climate data from approximately 2000-2024. These models are created as an ensemble of up to 24 models using 6 different modeling algorithms. Each of the 24 models was evaluated for its predictive ability, and poor models were excluded leading to each map on average being based on 14 models. These models are not perfect. There are necessarily many populations which exist outside of areas the models predict, as well as areas the models predict as suitable but are actually not.
Data Sources & cleaning. Data featured on this page were obtained from herbarium specimens, iNaturalist.org, roadside surveys by the author, occurrence data from the Invasive Species Committees, and vegetation plot data from the National Parks Inventory & Monitoring program, Hawaii forest bird survey, and the Ko'olau Mountain Watershed Partnership. To clean the data, records with coordinate uncertainty greater than 1000 meters were removed. The points were manually checked to ensure the coordinates were in reasonable areas and remove any points suspected to be misidentified. Points were also checked against rainfall and temperature from Hawaiʻi climate data portal to identify outliers or points from irrigated areas.
Many known populations are not yet mapped. The images below help show the spatial biases of the data. There are approximately 20000 herbarium specimens of grasses which were collected across Hawaiʻi, but the majority of these predate the GPS era and do not yet have coordinates associated with them. Approximately 5000 of these points are included, as various projects have added coordinates to these older specimens. Some of the efforts to add coordinates to these specimens have been of higher quality than others and many of the erroneous coordinates which were excluded from these maps came from such efforts. Future versions of this atlas will focus on mapping more of these records. Several thousand points have already been added since version 1.0.
Density of grass occurence data, note the poor sampling on Hawaiʻi and in areas far from roads and cities.
Richness of grasses (number of species recorded per area), note the higher richness around urban areas.
About names: Synonyms and common names are shown for each species, the common names are specifically those which are used in Hawaiʻi. For species which do not have widely uses common names here, english names from other regions are included. Synonyms are principally those which have been used historically in Hawaiʻi. Misapplied names are those which have been based on misidentifcations by past botantists due to the difficulties of identifying introduced grasses in Hawaiʻi.
Points from irrigated areas were removed.The maps show where grasses occur under natural climate conditions. As such, occurrence points of grasses found under irrigated conditions where the climate is otherwise too dry for their growth have been removed. Cultivated occurrences have also been removed when their description mentions them occurring under cultivation, but the cultivation status of some species in the ISC data were not always clear.
About island statuses. Islands are colored by whether each species currently occurs on it. If the species has been observed in the past 50 years it is considered present on the island. If it has not been seen in over 50 years it is considered likely extirpated ("extirpated?" on map legends) and if it has not been seen for a longer period on an island it is considered extirpated. Only one grass (Eragrostis mauiensis is truly extinct. Some are "questionable naturalizations" when it is not clear that the grass has formed a wild population or lack an herbarium specimen. "Waif" occurrences are largely of grain species which keep reappearing from birdseed or spilled grain but do not form sustaining populations.
Also see my other grass resources here.
Copyright Kevin Faccenda. Updated July 2024
synonyms: Echinochloa crus-pavonis [misapplied in Hawai'i], Panicum crus-galli, Echinochloa crus-galli var. crus-pavonis [misapplied in Hawai'i], Echinochloa walteri [misapplied in Hawai'i]
Density of grass occurence data, note the poor sampling on Hawaiʻi and in areas far from roads and cities.
Richness of grasses (number of species recorded per area), note the higher richness around urban areas.
