‘Integronsters’, integral and integrated modeling Voinov, Alexey; Shugart, Herman H.
Environmental modelling & software : with environment data news,
January 2013, 2013, 2013-01-00, Letnik:
39
Journal Article
Recenzirano
In many cases model integration treats models as software components only, ignoring the fluid relationship between models and reality, the evolving nature of models and their constant modification ...and recalibration. As a result, with integrated models we find increased complexity, where changes that used to impact only relatively contained models of subsystems, now propagate throughout the whole integrated system. This makes it harder to keep the overall complexity under control and, in a way, defeats the purpose of modularity, when efficiency is supposed to be gained from independent development of modules. Treating models only as software in solving the integration challenge may give birth to ‘integronsters’ – constructs that are perfectly valid as software products but ugly or even useless as models. We argue that one possible remedy is to learn to use data sets as modules and integrate them into the models. Then the data that are available for module calibration can serve as an intermediate linkage tool, sitting between modules and providing a module-independent baseline dynamics, which is then incremented when scenarios are to be run. In this case it is not the model output that is directed into the next model input, but model output is presented as a variation around the baseline trajectory, and it is this variation that is then fed into the next module down the chain. However still with growing overall complexity, calibration can become an important limiting factor, giving more promise to the integral approach, when the system is modeled and simplified as a whole.
Forests provide important ecosystem services such as carbon sequestration. Forest landscapes are intrinsically heterogeneous—a problem for biomass and productivity assessment using remote sensing. ...Forest structure constitutes valuable additional information for the improved estimation of these variables. However, survey of forest structure by remote sensing remains a challenge which results mainly from the differences in forest structure metrics derived by using remote sensing compared to classical structural metrics from field data. To understand these differences, remote sensing measurements were linked with an individual-based forest model. Forest structure was analyzed by lidar remote sensing using metrics for the horizontal and vertical structures. To investigate the role of forest structure for biomass and productivity estimations in temperate forests, 25 lidar metrics of 375,000 simulated forest stands were analyzed. For the lidar-based metrics, top-of-canopy height arose as the best predictor for describing horizontal forest structure. The standard deviation of the vertical foliage profile was the best predictor for the vertical heterogeneity of a forest. Forest structure was also an important factor for the determination of forest biomass and aboveground wood productivity. In particular, horizontal structure was essential for forest biomass estimation. Predicting aboveground wood productivity must take into account both horizontal and vertical structures. In a case study based on these findings, forest structure, biomass and aboveground wood productivity are mapped for whole of Germany. The dominant type of forest in Germany is dense but less vertically structured forest stands. The total biomass of all German forests is 2.3 Gt, and the total aboveground woody productivity is 43 Mt/year. Future remote sensing missions will have the capability to provide information on forest structure (e.g., from lidar or radar). This will lead to more accurate assessments of forest biomass and productivity. These estimations can be used to evaluate forest ecosystems related to climate regulation and biodiversity protection.
Few old-growth stands remain in the matrix of secondary forests that dominates the eastern North American landscape. These remnant stands offer insight on the potential carbon (C) storage capacity of ...now-recovering secondary forests. We surveyed the remaining old-growth forests on sites characteristic of the general Mid-Atlantic United States and estimated the size of multiple components of forest C storage. Within and between old-growth stands, variability in C density is high and related to overstory tree species composition. The sites contain 219 ± 46 Mg C/ha (mean ± SD), including live and dead aboveground biomass, leaf litter, and the soil O horizon, with over 20% stored in downed wood and snags. Stands dominated by tulip poplar (
Liriodendron tulipifera
) store the most live biomass, while the mixed oak (
Quercus
spp.) stands overall store more dead wood. Total C density is 30% higher (154 Mg C/ha), and dead wood C density is 1800% higher (46 Mg C/ha) in the old-growth forests than in the surrounding younger forests (120 and 5 Mg C/ha, respectively). The high density of dead wood in old growth relative to secondary forests reflects a stark difference in historical land use and, possibly, the legacy of the local disturbance (e.g., disease) history. Our results demonstrate the potential for dead wood to maintain the sink capacity of secondary forests for many decades to come.
Forest ecosystems are an important sink for terrestrial carbon sequestration. Hence, accurate modeling of the intra‐ and interannual variability of forest photosynthetic productivity remains a key ...objective in global biology. Applying climate‐driven leaf phenology and growth in models may improve predictions of the forest gross primary productivity (GPP). We used a dynamic non‐structural carbohydrates (NSC) model (FORCCHN2) that couples leaf development and phenology to investigate the relationships among photosynthesis and environmental factors. FORCCHN2 simulates spring and autumn phenological events from heat and chilling, respectively. Leaf area index data from satellites along with climate data estimated localized phenological parameters. NSC limitation, immediate temperature, accumulated heat, and growth potential comprised a daily leaf‐growth model. Functionally, leaf growth was decoupled from photosynthesis. Leaf biomass determined overall photosynthetic production. We compared this model with outputs of the other six terrestrial biospheric models and with observations from the North American Carbon Program Site Interim Synthesis in 18 forest sites. This model improved the predicted performance of yearly GPP with a 57%–210% increase in correlation (median) and up to a 102% reduction in biases (median), compared to three prognostic models and three prescribed models. At the North America continental scale, the model predicted the average annual GPP of 7.38 Pg C/year from forest ecosystems during 1985–2016. The results showed an increasing trend of GPP in North America (1.0 Pg C/decade). The inclusion of climate‐driven phenology and growth has a significant potential for improving dynamic vegetation models, and promotes a further understanding of the complex relationship between environment and photosynthesis.
This work developed a vegetation dynamic model (FORCCHN2 model) that couples leaf development and phenology to improve predictions of the forest gross primary productivity (GPP) in the North America forests. The model predicted the average annual GPP of 7.38 Pg C/year from forest ecosystems during 1985–2016 at the North America continental scale.
Soil–atmosphere exchange significantly influences the global atmospheric abundances of carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O). These greenhouse gases (GHGs) have been ...extensively studied at the soil profile level and extrapolated to coarser scales (regional and global). However, finer scale studies of soil aggregation have not received much attention, even though elucidating the GHG activities at the full spectrum of scales rather than just coarse levels is essential for reducing the large uncertainties in the current atmospheric budgets of these gases. Through synthesizing relevant studies, we propose that aggregates, as relatively separate micro‐environments embedded in a complex soil matrix, can be viewed as biogeochemical reactors of GHGs. Aggregate reactivity is determined by both aggregate size (which determines the reactor size) and the bulk soil environment including both biotic and abiotic factors (which further influence the reaction conditions). With a systematic, dynamic view of the soil system, implications of aggregate reactors for soil–atmosphere GHG exchange are determined by both an individual reactor's reactivity and dynamics in aggregate size distributions. Emerging evidence supports the contention that aggregate reactors significantly influence soil–atmosphere GHG exchange and may have global implications for carbon and nitrogen cycling. In the context of increasingly frequent and severe disturbances, we advocate more analyses of GHG activities at the aggregate scale. To complement data on aggregate reactors, we suggest developing bottom‐up aggregate‐based models (ABMs) that apply a trait‐based approach and incorporate soil system heterogeneity.
How should we view soils beneath our feet of huge heterogeneity? We argue for a bottom‐up approach that aggregates, as relatively separate micro‐environments embedded in a complex soil matrix, can be viewed as biogeochemical reactors of GHGs. Being inspired by the individual‐based models in ecological systems, especially forest systems since the 1970s, we further propose to develop aggregate‐based models (ABMs) to simulate soil systems composed of aggregates of different sizes.
"Where were you when I laid the foundation of the earth?" God asks Job in the "Whirlwind Speech," but Job cannot reply. This passage—which some environmentalists and religious scholars treat as a ..."green" creation myth—drives renowned ecologist H. H. Shugart's extraordinary investigation, in which he uses verses from God's speech to Job to explore the planetary system, animal domestication, sea-level rise, evolution, biodiversity, weather phenomena, and climate change. Shugart calls attention to the rich resonance between the Earth's natural history and the workings of religious feeling, the wisdom of biblical scripture, and the arguments of Bible ethicists. The divine questions that frame his study are quintessentially religious, and the global changes humans have wrought on the Earth operate not only in the physical, chemical, and biological spheres but also in the spiritual realm. Shugart offers a universal framework for recognizing and confronting the global challenges humans now face: the relationship between human technology and large-scale environmental degradation, the effect of invasive species on the integrity of ecosystems, the role of humans in generating wide biotic extinctions, and the future of our oceans and tides.
Global estimates of forest aboveground biomass and carbon storage have major discrepancies linked to limitations in tree-level biomass estimates. Robust allometric equations can improve biomass ...estimates; however, destructive sampling to measure single-tree biomass is expensive, challenging, and prone to measurement error. We present a method to efficiently and non-destructively estimate single-tree biomass from terrestrial LiDAR scan data and test the approach on 21 destructively-sampled lodgepole pine (Pinus contorta) trees. The approach estimates branch and foliage volume using voxelization and estimates trunk volume using a method developed in this study called the Outer Hull Model (OHM). The OHM iteratively fits convex hulls, accurately handles noisy scan data, and fits the true shape of the trunk rather than forcing a cylindrical fit. Volume from the LiDAR scans is converted to biomass using density values from the literature and from field sampling to assess model sensitivity to density values. Whole-tree aboveground biomass estimates derived from the LiDAR scans were nearly unbiased and agreed strongly with destructive sampling data (R2=0.98, RMSE=20.4kg). Estimation of the trunk component biomass (R2=0.99, RMSE=12.3kg) was stronger than foliage and needle component estimates (R2=0.54, RMSE=21.4kg). The approach presented in this study accurately and non-destructively estimated the aboveground biomass of needleleaf trees with minimal user input. The promising performance on coniferous trees advances efficient sampling of single-tree biomass.
•A novel algorithm is presented to estimate individual tree biomass from TLS data.•The algorithm uses the shape of the tree rather than forcing a cylinder to fit.•TLS biomass estimates were validated with destructive sampling data.•Estimates of whole-tree and trunk biomass were accurate.•The method advances efficient single-tree biomass estimation, improving allometry.
Air quality is closely associated with climate change via the biosphere because plants release large quantities of volatile organic compounds (VOC) that mediate both gaseous pollutants and aerosol ...dynamics. Earlier studies, which considered only leaf physiology and simply scale up from leaf-level enhancements of emissions, suggest that climate warming enhances whole forest VOC emissions, and these increased VOC emissions aggravate ozone pollution and secondary organic aerosol formation. Using an individual-based forest VOC emissions model, UVAFME-VOC, that simulates system-level emissions by explicitly simulating forest community dynamics to the individual tree level, ecological competition among the individuals of differing size and age, and radiative transfer and leaf function through the canopy, we find that climate warming only sometimes stimulates isoprene emissions (the single largest source of non-methane hydrocarbon) in a southeastern U.S. forest. These complex patterns result from the combination of higher temperatures’ stimulating emissions at the leaf level but decreasing the abundance of isoprene-emitting taxa at the community level by causing a decline in the abundance of isoprene-emitting species (Quercus spp.). This ecological effect eventually outweighs the physiological one, thus reducing overall emissions. Such reduced emissions have far-reaching implications for the climate–air-quality relationships that have been established on the paradigm of warming-enhancement VOC emissions from vegetation. This local scale modeling study suggests that community ecology rather than only individual physiology should be integrated into future studies of biosphere–climate–chemistry interactions.
Individual-based models (IBMs) of complex systems emerged in the 1960s and early 1970s, across diverse disciplines from astronomy to zoology. Ecological IBMs arose with seemingly independent origins ...out of the tradition of understanding the ecosystems dynamics of ecosystems from a 'bottom-up' accounting of the interactions of the parts. Individual trees are principal among the parts of forests. Because these models are computationally demanding, they have prospered as the power of digital computers has increased exponentially over the decades following the 1970s. This review will focus on a class of forest IBMs called gap models. Gap models simulate the changes in forests by simulating the birth, growth and death of each individual tree on a small plot of land. The summation of these plots comprise a forest (or set of sample plots on a forested landscape or region). Other, more aggregated forest IBMs have been used in global applications including cohort-based models, ecosystem demography models, etc. Gap models have been used to provide the parameters for these bulk models. Currently, gap models have grown from local-scale to continental-scale and even global-scale applications to assess the potential consequences of climate change on natural forests. Modifications to the models have enabled simulation of disturbances including fire, insect outbreak and harvest. Our objective in this review is to provide the reader with an overview of the history, motivation and applications, including theoretical applications, of these models. In a time of concern over global changes, gap models are essential tools to understand forest responses to climate change, modified disturbance regimes and other change agents. Development of forest surveys to provide the starting points for simulations and better estimates of the behavior of the diversity of tree species in response to the environment are continuing needs for improvement for these and other IBMs.