PROJECT SUMMARY
Enhancing the accuracy of energy/carbon flux estimates at all scales is a critical part of improving understanding of the interactions between land surface biospheric processes and global climate change. Current approaches that scale between regional estimates with data from remote sensing, eddy covariance flux, and intensive plant- and stand-level flux measurements assume estimates from these extremely small areas are representative of larger regions. The timing of leaf senescence (coloring and subsequent fall; i.e., phenology) during autumn has large impacts on lower atmospheric energy-mass exchange through differential carbon assimilation and transpiration totals across the landscape, which are equal to or greater than those of spring phenology. However, spatial variations in autumn phenological timing at the community level have not been systematically measured and analyzed, and underlying environmental drivers are not well understood. If large, autumn leaf senescence variations may reflect gradients in plant growth that could foster systematic errors in seasonal fluxes of equal or greater magnitude than those during earlier portions of the growing season. Thus, autumn phenological data collected in a spatially explicit manner offer considerable opportunities for gauging landscape-level spatial variations crucial for accurate scaling-up of flux measurements to larger areas or downscaling regional-scale atmospheric circulation models. In this project, spatial variability of autumn phenological data will be measured and analyzed at the community level, compared to microclimatic and remote sensing measurements, and used as the basis for regional-scale multi-species phenological models, which could contribute to increased accuracy of energy/carbon flux estimates across large areas.
INTELLECTUAL MERIT
This project addresses issues that are significant for advancement across the fields of climatology, plant physiology, ecology, and remote sensing. The nature of autumn phenological variability in space and time has not been previously recorded over a large area and combined with supporting measurements. Results from recent studies strongly suggest that understanding stand-level spatial patterns of autumn plant phenological development (and the environmental processes that drive them, especially when combined with information derived from spring phenological measurements in previous work), will provide the key knowledge needed to improve landscape level estimates of evapotranspiration and carbon accumulation across the entire growing season, as derived from moderate resolution remote sensing data. These results will in turn contribute to broader understanding of landscape variability, atmosphere-biosphere interactions, and the type of measurements that are necessary to accurately scale-up flux measurements to regional and continental areas.
BROADER IMPACTS
College-level students will be closely involved with the tasks in this project. One or more will use the data to support development of their doctoral dissertations. Two will collect (and one analyze) phenological data during each annual autumn field campaign. My research team has established initial cooperative agreements with two charter K-12 schools in northern Michigan (one focusing on American Indian culture). Students will take local phenological observations (as part of their science classes) using the USA-NPN site. We will help them understand the significance of their efforts by providing in-person interaction, and instruments to record temperature measurements at their school sites, thus better connecting them to the project’s scientific objectives. The detailed spatial measurements and analyses that proceed from this study will lay the foundation for future work that explicitly links phenological variations with plant physiological responses. The spatially concentrated phenological measures produced by this study will provide future near-surface and satellite-derived remote sensing studies with a record of autumn plant development and growth (complementary to already-collected spring phenology data) that contains vastly more information about species differences, spatial variability, and precise event timing that has typically been recorded in the past. These measures will present new opportunities for comparison and validation, and will inform phenological monitoring schemes at additional sites. Overall, the results of this project will also contribute to better understanding of the impacts of climate change on the biosphere, which will increase knowledge of potential future changes, and may allow for better planning relative to societal impacts of biospheric changes.
RESEARCH FINDINGS
This project focused on questions and analyses related to scaling-up of autumn tree phenological observations for proper comparison to satellite-derived (MODIS) measures, and comparison of autumn tree phenology to carbon flux measurements from the nearby WLEF/Park Falls Ameriflux tall tower. These areas were deemed most immediately useful for the broader scientific community, and most helpful to address remaining concerns regarding the relationship and significance of phenological measures compared to carbon flux changes observed during the end of the growing season.
This study evaluated the ability of MODIS Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI) based land surface phenology to track ground-based autumn landscape phenology (LP, based on visually observations) in a temperate mixed forest. In particular, MODIS VI-based dormancy onset was able to estimate LP full leaf coloration with a difference of 0-12 days; and VI time series closely tracked the progression of leaf coloration. In addition, minimum NDVI values at later autumn phenological developmental stages appeared to correspond with the time of full leaf fall. These findings support the usefulness of MODIS data for monitoring autumn phenology in a mixed forest. Similar studies replicated in other vegetated environments are needed for more comprehensive validation of remotely sensed autumn phenology. This study suggests an approach for building more reliable phenological monitoring systems by linking remote sensing and ground observations during the crucial, but less studied autumn season.
The main goal of these investigations is to better understand the annual variability among traditional (ground-based visual–recorded by a human observer) autumn phenological measurements, their environmental drivers, the differences among tree species, and how these measurements related to under-canopy light sensor measurements of leaf fall, as well as eddy covariance measurements of the seasonal change in carbon flux (net ecosystem exchange, NEE). Tree phenological observations and associated environmental measurements have been recorded in Downer Woods (an approximately 11 acre woodlot on the UW-Milwaukee campus, located at 43.081°N, 87.881°W) since 2007. The results show the considerable variation in the progression of annual leaf coloring between the two major species that compose Downer Woods, White Ash and Basswood. The confounding issues regarding the environmental drivers of leaf coloring and fall in autumn are also shown in these figures, by comparing all years, but especially 2007 and 2008. Across most years, the variation in White Ash is less than for Basswood, suggesting that White Ash coloring may be more controlled by day length/light levels and Basswood coloring more controlled by temperature. In 2007 and to a lesser extent in 2014, that seems be the case. However in 2008, the coloring progression of both tree species is nearly identical in timing and progression speed, suggesting a more complex relationship. These data will form the basis for additional study.
The work in Downer Woods has served as a convenient “test bed” for trying out and improving on techniques to apply in the more remote and difficult to access, but much larger and more diverse study site in northern Wisconsin (WLEF/Park Falls, 45.946°N, 90.272°W), located near the eddy covariance flux tower. We have had outstanding success relating under-canopy light sensor measurements to leaf phenological development in trees during spring, but more limited success in autumn. This appears to be primarily due to the more gradual growth process and long period over which leaves develop in spring, which provides a lot more “signal” for the sensors to recognize. In contrast during autumn, the simple sensors we use can’t detect leaf color, and can only detect leaf fall, which is a considerably more abrupt process that also occurs over a much shorter periods of time.
Nevertheless, we are having some success in relating light sensor measurements to visually recorded autumn phenology, and slowly understanding what is possible. There is a reasonably good correlation of under-canopy light sensor measurements to Basswood 50% (950) leaf fall dates in Downer Woods across seven years. There are also good correlations among light sensor measurements, Sugar Maple 90% (990) leaf fall dates, and an End of Autumn/Fall date derived for the NEE (carbon flux measurements) at the WLEF/Park Falls study site over six years (the light sensor was damaged in 2009, so no data). Logistical curve functions are used to transform the under-canopy light interception data and carbon flux (NEE) data streams into three dates, marking the transition periods in the season, start of fall (SOF), middle of fall (MOF), and end of fall (EOF). These dates can then be compared to other measurements.
Overall, the study has been successfully in achieving useful results in the areas outlined for emphasis by the reduced scope of work, and much foundation for on-going investigations into autumn phenological measurements and relationships.
OUTCOMES
Phenology is the study of recurring plant and animal life cycle stages (i.e., first spring leaves, fall leaf coloring, or return of migrating birds), especially their timing and relationships with weather and climate. Studying these changes in plants during spring and autumn improves understanding of the relationships among plant growth, temperatures, and other environmental factors that affect growing season length. Such information contributes to better understanding of potential climate change (atmospheric) impacts on the biosphere (living things), which will increase knowledge, and may allow for better societal planning.
However, complicating this goal, plant phenology is measured in a variety of ways, each with particular advantages. Combinations are beneficial, but many of the measurement types are not readily comparable without effort. This project expanded on previous successful work (conducted near the WLEF/Park Falls Ameriflux tall tower, 45.946°N, 90.272°W) to characterize the connections among satellite-derived measures (which combine information over large areas), traditional surface data (systematically recorded by ground-based observers for large numbers of individual trees), and instrumental measurements of carbon flux (which indicate composite photosynthesis-driven responses of all plants) during spring leaf development to the autumn leaf senescence period, thus allowing characterization of these interrelationships at both ends of the growing season. Spatial variations in ground-based autumn phenological timing (leaf coloring and leaf drop) have not previously been systematically measured and analyzed, and underlying environmental drivers (temperature, day length, moisture stress, etc.) are not well understood.
The results of this project demonstrated that satellite-derived measurements of phenology (MODIS sensor-based vegetation indices) are able to estimate ground-based autumn observations of tree full leaf coloration and full leaf fall timing successfully. These findings support the usefulness of MODIS data for monitoring autumn phenology in mixed forests. Similar studies replicated in other vegetated environments are needed for more comprehensive validation of satellite-derived autumn phenology. The results also showed the considerable variations in the progression of annual leaf coloring/fall among major tree species, and the confounding issues regarding environmental drivers of leaf coloring/fall in autumn. Lastly, the results showed progress in understanding how to utilize techniques that allow good correlations among concurrent acquisitions of under-canopy light sensor measurements, visually recorded (ground-based) autumn tree phenology, and end of Autumn/Fall dates derived from eddy covariance system carbon flux measurements.
Overall, this study suggests an effective approach for building more reliable phenological monitoring systems by linking remote sensing and ground-based observations during the crucial, but less studied autumn season, and provides the foundations for on-going and future investigations into autumn phenological measurements and relationships.