Between 1954 and the mid-1980s, about 50,000 ha of native montane rainforest on the island of Hawai‘i experienced a decline in canopy trees (“‘ōhi‘a dieback”), leading to great concern about the future of Hawai‘i’s rainforests. Dieback symptoms particularly affected the dominant tree species, the endemic Metrosideros polymorpha. Early hypotheses postulated that the forest decline was caused by a virulent pathogen or a combination of biotic disease and pest agents. This was ruled out after a decade of intensive disease research in the 1970s. Instead, it turned out that dieback patterns were significantly related to the physical environment, particularly the slope, topography, relative position on the hill slope, annual rainfall, and the type of substrate. Thus, an alternative hypothesis proposed that dieback is initiated by climate anomalies that manifest through soil moisture regimes under certain conditions of forest stand demography. Ironically, scientific perception of this interdisciplinary groundbreaking research that stimulated a global perspective on forest decline vanished while the awareness of climate change and its potential impact on the world’s forests started to grow, rapidly becoming a major focus of research in recent years. In this paper, we reinforce memory of the world’s first complex discussion on the natural causes of forest dieback as a showcase for the complexity of modern forest mortality research. This case demonstrates the need to rigorously identify, quantify, and fully understand all drivers of tree mortality to realistically project future climate-driven and other risks to forest ecosystem functions and services. Moreover, we summarize recent findings on forest mortality and climate change in the Pacific islands and beyond.

The Leeward Kohala Field System (LKFS) on Hawai‘i Island once featured over 60 km² of productive, rain-fed croplands. For several centuries, its occupants cultivated ‘uala (Ipomoea batatas, sweet potato) as a staple crop between kuaiwi (earthen or rock walls) planted with kō (Saccharum officinarum, sugarcane). These raised kō rows could have influenced ‘uala growth through an array of microclimatic processes, including wind abatement, shading, and the redistribution of moisture. While such effects are frequently mentioned in the literature, efforts to directly quantify them and relate them to ‘uala production have been lacking. We measured wind speeds, precipitation rates, solar illuminance, soil moisture, and ‘uala yields along a transect through three kō rows within the LKFS. Kō rows proved effective windbreaks, reducing near-surface wind speeds by up to 90% and for distances of up to 10 m. The rows also concentrated wind-blown moisture at their upwind edges while creating rain shadows 5–6 m in length. ‘Uala yields peaked at 3.8–6.0 kg m^–2 near the center of inter-row space, probably because ‘uala were energy-limited during the wet study period and grew slowly when shaded by the kō. A zone of low turbulence leeward of each kō row also appeared to benefit ‘uala growth. Additional measurements are needed to investigate the landscape-level hydrologic effects of kō row planting. Our findings will help guide ongoing efforts to expand agricultural and educational activities in the LKFS.

Metrosideros polymorpha (‘ōhi‘a, ‘ōhi‘a lehua) is an important foundation species in Hawaiian forest habitats. The genus originated in New Zealand and was dispersed to the Hawaiian archipelago approximately 3.9 million years ago. It evolved into five distinct endemic species and one of these, Metrosideros polymorpha, further differentiated into eight varieties across what are now the main Hawaiian Islands. ‘Ōhi‘a is a tree that has great significance in indigenous Hawaiian culture. It is considered a physical manifestation of several principal Hawaiian deities, and serves a broad range of uses in Hawaiian material culture. It occupies a wide diversity of habitats, extending from sea level to over 2,200 m elevation, occupying habitats that range from extremely wet to dry rainfall zones. It is the dominant or co-dominant tree species in wet and mesic forests and is also one of the first woody species to become established on young lava flows. Although ‘ōhi‘a is a dominant forest tree it also exhibits many characteristics of a pioneer species. ‘Ōhi‘a provides the matrix for a wide diversity of endemic plants and animals found in these habitats and functions as the primary vegetation cover on native Hawaiian watersheds, facilitating groundwater recharge and regulating surface runoff. ‘Ōhi‘a has shown remarkable resilience by recolonizing forests that were opened up by disturbance, such as the widespread ‘ōhi‘a canopy dieback that occurred on East Maui in the 1900s and on the east side of the Island of Hawai‘i in the 1970s. Several human-related conditions threaten the continued stability of Hawaii’s native ecosystems, including invasive plants, plant diseases, introduced animals, and changing climate. The research and conservation legacy of Dr. Dieter Mueller-Dombois helped to expand our knowledge of the ecology and importance of ‘ōhi‘a forests, and to increase awareness and appreciation of the remarkable Hawaiian ecosystems that are unique to the world.

Between 1958 and 2018, multiple roadside surveys were conducted on the upper altitudinal limits of the alpine Maunaloa (now preferred Hawaiian language spelling of Mauna Loa) vascular flora from near tree line (2,525 m) to the NOAA Mauna Loa Observatory (MLO) at 3,397 m. Five native plant species were encountered in 1958 on numerous sparsely vegetated historic and prehistoric basaltic lava flows. A resurvey of the roadside 50 years later (2008) yielded 22 species including nine new native species and eight aliens. The aliens were limited to a few individuals at sites disturbed by human activity. Here, floristic change along the transect was reassessed after 60 years (2018), recording a total of 30 species with 15 natives and 15 aliens. Floristic diversity and shifts in altitudinal limits are evaluated in relation to well documented and dramatic local climate change over recent decades, along with the competing influences of substrate-control and increasing human disturbance. The results indicate that lava substrate age and textural variation, along with human mediated propagule pressure, and associated disturbance exert greater impact than rapid climate change in explaining current patterns of plant distribution and diversity in this hyper-arid, alpine environment. This long-term monitoring effort of alpine lava flows has also informed and reinforced a more general moisture/ substrate-control model for early primary succession extending to lower subalpine mountain elevations involving the two dominant lava flow textural classes: ‘a‘ā and pāhoehoe.

Tropical cyclones are common disturbances on many Pacific islands and affect forest structure and community composition. However, we know little about the process of succession after disturbances in the tropical South Pacific. We utilize published data of vegetation surveys in a lowland tropical rainforest reserve in Samoa that were undertaken within 1 year, and after 6 and 15 years, after two cyclones and a fire that occurred in 1990 and 1991. We combine these surveys with functional trait data from literature and the field. Community and functional composition differed little from inferred pre-cyclone conditions soon after the disturbances but had changed considerably 6 years after the disturbance. Early successional species with functional trait characteristics relating to resources acquisition and faster growth (lower wood density, larger leaf area, shorter maximum height, smaller seeds) had become dominant 6 years after the disturbances, but had declined considerably by 15 years. No clear differences in community-weighted means of functional traits were detected between burned and unburned forest, but community composition differed considerably. In particular, the introduced rubber tree, Funtumia elastica, which was functionally intermediate between early and late successional species, had become very abundant in burned forest. Our results suggest that ecological functions may be more resilient to cyclone disturbance than community composition, but this requires further study. Our findings highlight the impact of cyclones on community composition and functioning, the importance of long-term data for investigating the recovery after disturbances, and the potential of multiple disturbances to facilitate the proliferation of invasive species.

Littoral and swamp forests are among the most threatened native plant communities in the island of Tahiti (South Pacific) due to past and present anthropogenic pressures such as agriculture, urbanization, pollutions, and invasive alien species, including the mangrove tree Rhizophora stylosa. In order to provide reference data that are crucial to implement appropriate conservation and restoration strategies in these habitats, we assessed the composition and structure of seven littoral and swamp forests types in eighteen 10 × 20 m plots considering three different strata (trees and lianas, epiphytes, and understory). Forest types were compared using common diversity indices (e.g., Shannon index, Simpson index, and Pielou evenness) and indicator values. Results show that native submangrove swamp forests dominated by the tree Talipariti tiliaceum and the large erect fern Acrostichum aureum were the most species-rich, while introduced Rhizophora mangroves had an almost monospecific composition in the trees, lianas, and understory strata. The diversity of trees and lianas was higher in littoral forests with the highest understory cover in Talipariti-Barringtonia asiatica and Talipariti-Inocarpus fagifer plant communities. Surprisingly, epiphyte diversity and abundance were higher in swamp forests with lower canopy, especially in the mixed swamp forest with both A. aureum and R. stylosa. These counter-intuitive results highlight the potential role of introduced species in creating novel microhabitats suitable for the development of some native epiphytes. Conservation and restoration projects should however focus on the use of native species to maintain these remnant littoral and swamp habitats in Tahiti and other high volcanic islands of the Society archipelago, rather than nonnative and potentially harmful species such as R. stylosa.

The powerful El Niño of 1982–1983 precipitated a severe drought in the Hawai‘i Islands and was followed by an unusually dry La Niña year. Our 1983/1984 study of the early successional demography of five shrub and one tree species on volcanic cinder, Big Island of Hawai‘i, inadvertently coincided with the end of the ENSO drought. Life stage structure analyses showed a short-term dieback in the populations but then rapid population recovery. A new demographic tool, population flow diagram analysis, was developed as an aid to interpret the temporal dynamics of life stage structure. Crown size demographic depletion models were also used to describe the species’ vital statistics. The apparent dieback was shown to be a temporary dormancy response to the El Niño/La Niña-induced drought rather than a true case of dieback related to cohort senescence. As precipitation levels returned to normal the populations were rejuvenated by the revival of senescent and dormant individuals. The species showed robust demographic resilience to an unusually powerful drought. The populations of the Devastation Area appear to be members of a non-equilibrium community but there was evidence of a shift towards equilibrium. Climate change may intensify ENSO droughts in Hawai‘i and could cause longer-term diebacks of these populations and possibly their extirpation, affecting the rate and nature of primary succession on volcanic cinder ecosystems. Population viability modelling could determine if the species are likely to face extirpation from climate-change-driven alterations in the historic pattern of El Niño/La Niña events.

Globally, subalpine and alpine plant communities are receiving increasing attention owing to disproportionately rapid warming at high altitudes, and the resultant habitat shrinkage leaving high-altitude specialists with nowhere to migrate. The Hawaiian subalpine zone (1,700–3,000 m) is an interesting example of this potential phenomenon because of the high endemism. We analyzed plant species richness, cover, and density from 89 plots (1,000 m²) sampled in 2010–2018 across two volcanic mountains, Haleakalā on Maui, and Mauna Loa on Hawai‘i. Most of the 139 plant species recorded were non-native (55%), with the remainder endemic (31%) and indigenous (14%). Plot-level richness differed from gamma diversity, with endemic species more abundant than non-native species. Non-native species richness was higher on Haleakalā than Mauna Loa. These communities are patchy with low-lying (<1 m) vegetation, and lower cover on younger drier Mauna Loa (36%) than Haleakalā (54%). Density was largely consistent with the understory cover data, with endemic Vaccinium reticulatum (>3,500/ha) and indigenous Leptecophylla tameiameiae (>2,430/ha) shrubs dominant on both volcanoes. Woodland communities were encountered only on Mauna Loa, with endemic trees Metrosideros polymorpha on wetter, south aspects, and Sophora chrysophylla on the drier, leeward side. Hawaiian subalpine vegetation varies among islands, volcanoes, and aspects, yet remains largely native-dominated, though with increasing threats from climate change, invasive non-native species, and wildfire. We recommend continued monitoring of biotic communities and climate in this sensitive zone, in situ physiological studies for the native matrix species, stricter non-native species biosecurity and sanitation protocols, wildfire prevention, and improved documentation of the effects of feral ungulates.

An individual-based, multiple-growthform gap model of forest stand development and succession (ZELIG.MGF) of the JABOWA and FORET lineage was modified to simulate long-term changes in ‘ōhi'a-hāpu'u montane rain forests on the island of Hawai‘i. Based on the autecology, architecture, and life history of the two dominant species, we were able to re-create some of the stand dynamics and population structures observed in these forests. The phenomenon of displacement dieback, which occurs only on rich sites with persistent cloud cover, is portrayed as a natural successional consequence of ‘ōhi’a lehua (Metrosideros polymorpha Gaud.) senescence and shade intolerance, which contrasts with the shorter stature, clonal reproduction, and shade tolerance of hāpu'u pulu tree fern (Cibotium glaucum (Sm.) Hook. & Arn.). ZELIG.MGF predicts that ‘ōhi’a will achieve a maximum basal area of 27 m²/ha in stands 80–90 years old, after which basal area is projected to decline to levels of 8–11 m²/ha that persist after 220 years. Alternating phases of ‘ōhi'a and hāpu'u may dominate individual gaps, but overall old-growth ‘ōhi’a populations do not recover to earlier levels. ‘Ōhi’a overstory mortality is consistent with senescence or a growing imbalance of respiratory to photosynthetic tissue in large trees. Understory mortality as modeled is largely due to shading by adults and by hāpu’u tree ferns, although mechanical damage from dead hāpu’u fronds, which was not modeled, may also be important. ‘Ōhi’a stand rejuvenation can occur when the density of hāpu’u is reduced by harvesting or wind storms.

Severe disturbances of landscapes entail an ecosystem development with the formation of structures and functions which may reach either a new equilibrium state or a state similar to the original ecosystem. Natural disturbances can result from major events such as volcanoes, glaciers, or denudations from landslides. Major disturbance may also evolve from anthropogenic influences such as from mining operations. They all can be considered starting points for the development of ecosystems from ‘point zero’, which was one central research interest of Dieter Mueller-Dombois. In this paper results from research in the Lusatian post-mining landscapes (Eastern Germany) are presented. Different methodological approaches are discussed. The well-defined ‘point zero’ of the ecosystem development allows for research on chronosequence designs as well as real time series studies. Chronosequences have been investigated to gain insight into the medium to long-term direction of the development. Real time series are recorded to obtain a more detailed understanding. The paper is structured into three main parts: First, effects of ecosystem disturbances by mining in Lusatia and the initial conditions for restoration are presented. In the following part practical rehabilitation measures and land use options are discussed. Finally, the third part summarizes results of long-term monitoring in an artificial watershed. In conclusion, post-mining landscapes allow for relevant case studies of ecosystem development after severe disturbances. Particularly, the starting point of the initial phase is very well defined which distinguishes these anthropogenically disturbed landscapes from landscapes after natural disturbances.

This study evaluates the secondary successional pathways of Hawaiian lowland rainforest following anthropogenic disturbance. Whereas primary succession on lava flows has been well studied in Hawaii, secondary successional dynamics on disturbed habitats in the presence of nonnative species has been poorly researched and documented. Our study was based on two vegetation sampling efforts conducted on 200–400-year-old lava flows 2 and 27 years following a 400-hectare clearcutting operation in the Puna District of the southeast quadrant of Hawaii Island. Intact forest adjacent to the clearcut area provided a baseline to compare against succession dynamics in the clearcut area. The most important trees of the uncut primary forest, based on frequency, density, and dominance, were keystone Hawaiian forest species Metrosideros polymorpha, followed by Psychotria hawaiiensis, Psidium cattleianum, Ciboteum glaucum, and Cibotium chamissoi. Whereas early successional nonnative species dominated the initial vegetation colonizing the clearcut site, M. polymorpha, P. cattleianum, and Cibotium species were also present. Twenty-seven years later we found the clearcut site colonized by two distinct secondary forest types: M. polymorpha-dominated forest and Falcataria moluccana-dominated forest. Whereas M. polymorpha maintained a similar level of dominance in uncut primary forest and M. polymorpha-dominated secondary forest stands, it was nearly absent from F. moluccana-dominated secondary forest stands. Results demonstrated that the keystone Hawaiian native tree species can effectively reestablish across disturbed landscapes, but are unable to outcompete in areas where fast growing nonnative trees become established. This underscores the critical need to suppress early nonnative tree establishment in disturbed habitats if the objective is to reestablish Hawaiian native forests.

In the interview transcript below, Emeritus Professor Dr. Dieter Mueller-Dombois talks about his prolific career in the Department of Botany (now a program within the Life Sciences Department) at the University of Hawai‘i at Mānoa in Honolulu, Hawai‘i. Dr. Mueller-Dombois describes major projects and publications across his time in Canada, Sri Lanka, and Hawai‘i, his experiences as a soldier in WWII, and his deep appreciation for the Hawaiian flora and for the Botany Department that inspired and nourished his far-reaching and deeply impactful contributions to the field of forest ecology. Included is a discussion of his work on ‘ōhi‘a (Metrosideros polymorpha) stand-level dieback in the 1980s in connection to the more recent episode of rapid ‘ōhi‘a death. Although the Botany Department is now merged with other departments, Dr. Mueller-Dombois hoped for it to remain with a distinct identity. In his view, it was an amiable department that made critical contributions to understanding and protecting the tremendous natural and cultural heritage of Hawai‘i.