Browsing by Subject "remote sensing"

Now showing 1 - 13 of 13

Canopy nitrogen, carbon assimilation, and albedo in temperate and boreal forests: Functional relations and potential climate feedbacks
(National Academy of Sciences, 2008) Ollinger, S V; Richardson, A D; Martin, M E; Hollinger, D Y; Frolking, S E; Reich, Peter B; Plourdea, L C; Katul, G G; Munger, J W; Orend, R; Smith, M L; Paw U, K T; Bolstad, P V; Cook, B D; Day, M C; Martin, T A; Monson, R K; Schmid, H P
The availability of nitrogen represents a key constraint on carbon cycling in terrestrial ecosystems, and it is largely in this capacity that the role of N in the Earth’s climate system has been considered. Despite this, few studies have included continuous variation in plant N status as a driver of broad-scale carbon cycle analyses. This is partly because of uncertainties in how leaf-level physiological relationships scale to whole ecosystems and because methods for regional to continental detection of plant N concentrations have yet to be developed. Here, we show that ecosystem CO2 uptake capacity in temperate and boreal forests scales directly with whole-canopy N concentrations, mirroring a leaf-level trend that has been observed for woody plants worldwide. We further show that both CO2 uptake capacity and canopy N concentration are strongly and positively correlated with shortwave surface albedo. These results suggest that N plays an additional, and overlooked, role in the climate system via its influence on vegetation reflectivity and shortwave surface energy exchange. We also demonstrate that much of the spatial variation in canopy N can be detected by using broad-band satellite sensors, offering a means through which these findings can be applied toward improved application of coupled carbon cycle–climate models.
Combining Machine Learning with Computer Vision for Precision Agriculture Applications
(2018-04) Zermas, Dimitris
Financial and social elements of modern societies are closely connected to the cultivation of corn. Due to its massive production, deficiencies during the cultivation process directly translate to major financial losses. Existing field monitoring solutions utilize aerial and ground means towards identifying sectors of the farmland presenting under-performing crops. Nevertheless, an inference element is still absent; that is the automated diagnose of the cause and severity of the deficiency. The early detection and treatment of crops deficiencies and the frequent evaluation of their growth status are thus tasks of great significance. Towards an automated health condition assessment, this thesis introduces schemes for the computation of plant health indices. First, we propose a methodology to detect nitrogen (N) deficiencies in corn fields and assess their severity at an early stage using low-cost RGB sensors. The introduced methodology is twofold. First, a low complexity recommendation scheme identifies candidate plants exhibiting nitrogen deficiency and second, a detection elimination step completes the inference loop by deciding which of the candidate plants are actually exhibiting that condition. Experimental results on a diverse real-world dataset achieve a 90.6% accuracy for the detection of N-deficient regions and support the extension of this methodology to other crops and deficiencies that show similar visual characteristics. Second, based on the 3D reconstruction of small batches of corn plants at growth stages between ''V3'' and ''V6'', an automated alternative to existing manual and cumbersome phenotype estimation methodologies is presented. The use of 3D models provides an elevated information content, when compared to planar methods, mainly due to the alleviation of leaf occlusions. High-resolution images of corn stalks are collected and used to obtain 3D models of plants of interest. Based on the extracted 3D point clouds, the calculation of a plethora of phenotypic characteristics for each 3D reconstruction are obtained such as the number of plants depicted with 88.1% accuracy, Leaf Area Index (LAI) with 92.48% accuracy, the height with 89.2% accuracy, the leaf length with 74.8% accuracy, and the location and the angles of leaves with respect to the stem. The last two variables are connected by showing the trend of the angles to change with respect to the leaf position on the stem as the crops grow. An experimental validation using both artificially made corn plants emulating real-world scenarios and real corn plants in different growth stages supports the efficacy of the proposed methodology. Although the proposed methodologies are agnostic to the platform that performs the data collection, for the presented experiments a MikroKopter Okto XL equipped with a Nikon D7200 RGB sensor and a DJI Matrice 100 with a Zenmuse X3 and a Zenmuze Z3 RGB high-resolution cameras were used. The flight altitude ranged between 6 and 15 m and the resolution of the images varies within a range of 0.2 to 0.47 cm/pixel. Thorough data collection and interpretation leads to a better understanding of the needs not only of the farm as a whole but to each individual plant providing a much higher granularity to potential treatment strategies. Through the thoughtful utilization of modern computer vision techniques, it is possible to achieve positive financial and environmental results for these tasks. The conclusions of this work, suggest a fully automated scheme for information gathering in modern farms capable of replacing current labor-intensive procedures, thus greatly impacting the timely detection of crop deficiencies.
Establishing The Feasibility Of Making Fine-Scale Measurements Of Habitat Use By White-Tailed Deer In Northern Minnesota
(2020-01) Smith, Bradley
Advances in technology enhance our ability to understand wildlife-habitat relationships. The Minnesota Department of Natural Resources’ new statewide white-tailed deer (Odocoileus virginianus) management plan aims to enhance its ability to maintain regional deer numbers near population goals. Habitat management is acknowledged as a key component to achieving the plan’s objectives. Informed habitat management prescriptions, based on an improved understanding of optimal size, shape, and arrangement of forest stands and foraging sites, and edge relationships, will contribute to a more successful integration of long-term forest and deer habitat management strategies. The objectives of my study were to establish the feasibility of combining cutting-edge Global positioning system (GPS) collar, remote sensing, and Geographic Information System technologies to 1) classify and inventory available habitat on deer winter ranges and 2) characterize how deer use habitat at the stand level to facilitate an improved understanding of their habitat requirements in northern Minnesota. During winter 2017–2018, 20 adult female deer were captured and fitted with GPS collars on 2 study areas (10/site) in northcentral (Inguadona Lake [IN]) and northeastern (Elephant Lake [EL]) Minnesota, with an additional 40 collars (20/site) deployed during winter 2018–2019. Prior to the deployment of GPS collars on free-ranging deer, I conducted stationary tests to evaluate the location-fix-success and spatial accuracy of 48 collars placed in 4 different cover types. The overall mean location error of the GPS collars was 5.7 m (± 0.15, range = 0–189), with errors in dense conifer (10.3 ± 0.52, range = 0–189 m) being greater than in hardwood stands (6.2 ± 0.22, range = 0–91 m), browse patches (3.2 ± 0.08, range = 0–26 m), and openings (3.2 ± 0.08, range = 0–32 m). With incorporation into the collars of quick fix pseudoranging (QFP) programming, I recovered 100% of the location-fixes during the stationary tests and from 30 collars deployed on free-ranging deer. Spatially, dense conifer stands accounted for 21% and 9%, and moderately dense conifer stands for 4% and 10% of the EL and IN sites, respectively. The proportion of forage openings was 9% on both sites. The mean size (area) of available dense conifer stands was similar on both study sites (6.7, 95% CI = 4.94–8.54 ha vs 6.0, 95% CI = 4.68–7.23 ha). Available forest stands were generally circular, providing a larger core area and less edge, with a mean edge:area ratio <400 m/ha. Deer use of cover types was highly variable among individuals. Mean individual use of dense conifer stands was 23% (range = 0–79%) and 9% (range = 0–29%), and mean use of forage openings was 13% (range = 0–42%) and 24% (range = 0–70%) at the EL and IN sites, respectively. To better understand deer use at the stand level and the arrangement of cover types, I measured the distance from each location-fix to the nearest dense conifer stand and forage opening. While using forage openings, deer were a mean of 177 m (± 7, range = 0–833) and 195 m (± 4, range = 0–882) from dense conifer stands at EL and IN. Likewise, individuals using dense conifer stands were a mean of 241 m (± 6, range =0–777) and 147 m (± 8, range =0–1,030) from forage openings at the respective sites. The use of an integrated technological approach is essential to a more thorough understanding of seasonal habitat requirements of deer. The ability to retrieve 100% of location-fixes with high spatial accuracy will allow us to confidently assess winter habitat use by white-tailed deer as winter progresses and assist managers in formulating prescriptions that effectively integrate forest and habitat management strategies and activities.
The imprint of plants on ecosystem functioning: A data-driven approach
(Elsevier, 2015) Musavi, Talie; Mahecha, Miguel D; Migliavacca, Mirco; Reichstein, Markus; van de Weg, Martine Janet; van Bodegom, Peter M; Bahn, Michael; Wirth, Christian; Reich, Peter B; Schrodt, Franziska; Kattge, Jens
Terrestrial ecosystems strongly determine the exchange of carbon, water and energy between the biosphere and atmosphere. These exchanges are influenced by environmental conditions (e.g., local meteorology, soils), but generally mediated by organisms. Often, mathematical descriptions of these processes are implemented in terrestrial biosphere models. Model implementations of this kind should be evaluated by empirical analyses of relationships between observed patterns of ecosystem functioning, vegetation structure, plant traits, and environmental conditions. However, the question of how to describe the imprint of plants on ecosystem functioning based on observations has not yet been systematically investigated. One approach might be to identify and quantify functional attributes or responsiveness of ecosystems (often very short-term in nature) that contribute to the long-term (i.e., annual but also seasonal or daily) metrics commonly in use. Here we define these patterns as “ecosystem functional properties”, or EFPs. Such as the ecosystem capacity of carbon assimilation or the maximum light use efficiency of an ecosystem. While EFPs should be directly derivable from flux measurements at the ecosystem level, we posit that these inherently include the influence of specific plant traits and their local heterogeneity. We present different options of upscaling in situ measured plant traits to the ecosystem level (ecosystem vegetation properties – EVPs) and provide examples of empirical analyses on plants’ imprint on ecosystem functioning by combining in situ measured plant traits and ecosystem flux measurements. Finally, we discuss how recent advances in remote sensing contribute to this framework.
Incorporating Remote Sensing into Forest Carbon Accounting using Model-based and Model-assisted Estimators
(2022-05) Emick, Ethan
Forests’ ability to sequester carbon dioxide (CO2) from the atmosphere has been iden-tified as a solution to mitigate anthropogenically caused climate change. The impor- tant role forests play in storing carbon has elevated the need to quantify aboveground biomass (AGB) stocks. Conventional estimation of AGB often uses probabilistically sampled field data to make population estimates of AGB. Estimates with the desired level of precision from field-only inventories typically require an intensive sampling de- sign that is expensive to collect. A method to increase estimation precision or reduce the number of needed field sample plots is to incorporate remote sensing information into AGB inventories. Using both the design-based and model-based inferential paradigm, the relationship between observed AGB and remote sensing variables can be exploited to increase estimation precision compared to traditional field-only inventories. This thesis details two projects that leverage remote sensing data to improve estimation precision for AGB as well as other forest inventory variables. Project 1: Crowd-sourcing field and remote sensing datasets is an excellent way to accumulate the training information needed to create regional AGB maps using machine learning (ML). However, ML-based estimators can be biased when they are trained using remote sensing data collected with multiple different sensor systems and field data collected with multiple different sampling protocols. Here, we demonstrate how to correct for potential ML estimator bias using model-assisted (MA) and geostatistical- model-based (GMB) estimators. Using probabilistically sampled field AGB values from the USFS Forest Inventory and Analysis (FIA) Program as the response variable and ML predictions as the predictor variable, we show how MA and GMB estimators can correct for ML estimator bias and be used to generate statistically defensible uncertainty estimates. Our motivating dataset is a collection of AGB maps generated using Random Forest (RF-AGB) and a probability sample of AGB values for Oregon from 2001 to 2016. Using the maps generated from the RF-AGB estimator, we apply MA and GMB estimators to generate areal density estimates at the state and county level. We also explore a case study at the HJ Andrews Experimental Forest using the GMB estimator at smaller scales than the county level. Results show that the proposed AGB density estimators that leverage the RF-AGB predictions are more precise those based on field data alone. Project 2: Inclusion of remote sensing data into forest inventories has been pro- posed to help improve estimation precision of AGB and other forest inventory variables, a method known as an enhanced forest inventory. An under-represented remote sens- ing technology in enhanced forest inventories is terrestrial laser scanning (TLS). TLS data are collected from the ground, making it challenging to collect wall-to-wall for an entire forest. Sampling strategies can be used to collect TLS data for a representative subset of locations within the forest. In this study we make use of a model-assisted double sampling estimator to relate TLS data to field measured AGB, basal area and tree density. We also include a canopy height model derived using digital aerial pho- togrammetry and the normalized difference vegetation index to assess improvements in estimation precision compared to using field data alone.
Investigating Nitrogen and Irrigation Management Strategies to Improve Agronomic and Environmental Outcomes for Potato Production
(2021-08) Bohman, Brian
Nitrogen [N] fertilizer and irrigation management practices are both critical factors for determining agronomic and environmental outcomes for potato [Solanum tuberosum (L.)] production. This dissertation was comprised of two overall objectives. First, a small-plot experiment evaluating the effects of six N rate, source, and timing treatments and two irrigation rate treatments on tuber yield, quality, net profitability, nitrate leaching, residual soil nitrate, plant N uptake, N nutrition index [NNI], N uptake efficiency, N utilization efficiency [NUtE], N use efficiency [NUE], biomass, harvest index, biomass, and potential N losses for potato [cv. ‘Russet Burbank’] were investigated in 2016 and 2017 at Becker, MN, on a Hubbard loamy sand. Conventional N fertilizer best management practices [BMPs] (270 kg N ha-1) were compared to reduced N rate (180 kg N ha-1), control N rate (45 kg N ha-1), and a variable rate [VR] N treatment based on the N sufficiency index [NSI] approach using remote sensing. Irrigation treatments included a conventional rate (100%) based on the “checkbook” method and a reduced rate (85%). The VR treatment reduced N applied relative to the recommended rate by 22 and 44 kg N ha−1 in 2016 and 2017, respectively. Irrigation rate was reduced by 29 and 33 mm in 2016 and 2017, respectively. From an agronomic perspective, neither VR N nor reduced irrigation produced significant differences in tuber yield or net return compared to full rate treatments. From an environmental perspective, nitrate leaching losses varied between 2016 and 2017 with flow-weighted mean nitrate N concentrations of 5.6 and 12.8 mg N L−1, respectively, and increased from 7.1 to 10.4 mg N L−1 as N rate increased from 45 to 270 kg N ha−1. Despite reductions in N rate for the VR N treatment, there was no significant difference in nitrate leaching compared with the existing N best management practices (BMPs). However, reducing irrigation rate by 15% decreased nitrate leaching load by 17% through a reduction in percolation. Second, an evaluation of the relationship between NUE, NNI, and their variation across genotype [G] x environment [E] effects was conducted. A novel theoretical relationship between NNI and NUtE was derived: at a constant NNI value, NUtE values increased non-linearly as biomass increased, and at an NNI value of 1.0 this relationship defines the critical N utilization efficiency curve [CNUtEC]. Subsequently, an evaluation of the variation in critical N concentration [%Nc] was conducted using a hierarchical Bayesian framework to infer the critical N dilution curve [CNDC] across G x E effects observed from multiple experimental trials. This statistical method was able to quantify the uncertainty in %Nc, which was used to directly compare CNDCs. Critical N concentration was found to significantly vary across the effect of E, and in some cases for G within E. Therefore, consideration of both NNI and NUE require explicit consideration of the uncertainty in and variation due to G x E effects for %Nc. Overall, the findings of this dissertation improve both the empirical and theoretical understanding of the impact of N fertilizer and irrigation management practices on agronomic and environmental outcomes for potato.
Mechanistic links between physiology and spectral reflectance enable pre-visual detection of oak wilt and drought stress
(Proceedings of the National Academy of Sciences, 2024-02) Sapes, Gerard; Schroeder, Lucy; Scott, Allison; Clark, Isaiah; Juzwik, Jennifer; Montgomery, Rebecca; Guzmán Q., J. Antonio; Cavender-Bares, Jeannine
Tree mortality due to global change—including range expansion of invasive pests and pathogens—is a paramount threat to forest ecosystems. Oak forests are among the most prevalent and valuable ecosystems both ecologically and economically in the United States. There is increasing interest in monitoring oak decline and death due to both drought and the oak wilt pathogen (Bretziella fagacearum). We combined anatomical and ecophysiological measurements with spectroscopy at leaf, canopy, and airborne levels to enable differentiation of oak wilt and drought, and detection prior to visible symptom appearance. We performed an outdoor potted experiment with Quercus rubra saplings subjected to drought stress and/or artificially inoculated with the pathogen. Models developed from spectral reflectance accurately predicted ecophysiological indicators of oak wilt and drought decline in both potted and field experiments with naturally grown saplings. Both oak wilt and drought resulted in blocked water transport through xylem conduits. However, oak wilt impaired conduits in localized regions of the xylem due to formation of tyloses instead of emboli. The localized tylose formation resulted in more variable canopy photosynthesis and water content in diseased trees than drought-stressed ones. Reflectance signatures of plant photosynthesis, water content and cellular damage detected oak wilt and drought 13 days before visual symptoms appeared. Our results show that leaf spectral reflectance models predict ecophysiological processes relevant to detection and differentiation of disease and drought. Coupling spectral models that detect physiological change with spatial information enhances capacity to differentiate plant stress types such as oak wilt and drought.
Predicting agronomic performance of barley using canopy reflectance data
(Canadian Journal of Plant Pathology, 2004) Steffenson, Brian; Fetch, T.G.; Pederson, V.D.
The ability to accurately and rapidly predetermine agronomic performance would be desirable in most plant breeding programs. Remote sensing of canopy reflectance is a quick and nondestructive method that may be useful in the estimation of agronomic performance. Studies were conducted at Fargo and Langdon, North Dakota, to determine the effectiveness of a multispectral radiometer in estimating yield, kernel plumpness (KP), and 1000-kernel weight (TKW) in barley. Canopy reflectance was measured in eight (500–850 nm) discrete narrow-wavelength bands. Three types of reflectance models were evaluated: simple models using one to four wavelengths, simple ratio and normalized difference vegetation indices (NDVI) using green, red, and near-infrared wavelengths, and soil-adjusted vegetation indices (SAVI). The relationship between canopy reflectance and agronomic performance was significantly influenced by environment, growth stage, and plant genotype. Grain yield was best estimated near GS73 (0.84 < R2 < 0.92) at Fargo and at GS83 (0.55 < R2 < 0.81) at Langdon. In contrast, KP and TKW could be estimated at both late (GS83; 0.68 < R2 < 0.93) and early (GS24–GS47; 0.72 < R2 < 0.91) growth stages. The 550-nm and 800-nm wavelengths are critical for development of predictive models. A simple model using 550-nm, 600-nm, and 800-nm from GS47-GS73 gave significant (0.45 < R2 < 0.64) estimation of agronomic performance across all environments. In contrast, simple ratio, NDVI, and SAVI were less effective (0.05 < R2 < 0.77) in predicting agronomic performance. Remote sensing using canopy reflectance is a potential tool to estimate agronomic performance of barley, but genotypic and crop stage factors affect this method. Further studies are needed to improve the usefulness of multispectral radiometry in predicting agronomic performance.
Remotely detected aboveground plant function predicts belowground processes in two prairie diversity experiments
(2021-06-08) Cavender-Bares, Jeannine; Schweiger, Anna K.; Gamon, John; Gholizadeh, Hamed; Kimberly, Helzer; Lapadat, Cathleen; Madritch, Michael; Townsend, Philip A.; Wang, Zhihui; Hobbie, Sarah E.; cavender@umn.edu; Cavender-Bares, Jeannine
Imaging spectroscopy provides the opportunity to incorporate leaf and canopy optical data into ecological studies, but the extent to which remote sensing of vegetation can enhance the study of belowground processes is not well understood. In terrestrial systems, aboveground and belowground vegetation quantity and quality are coupled, and both influence belowground microbial processes and nutrient cycling, providing a potential link between remote sensing and belowground processes. We hypothesized that ecosystem productivity, and the chemical, structural and phylogenetic-functional composition of plant communities would be detectable with remote sensing and could be used to predict belowground plant and soil processes in two grassland biodiversity experiments—the BioDIV experiment at Cedar Creek Ecosystem Science Reserve in Minnesota and the Wood River Nature Conservancy experiment in Nebraska. Specifically, we tested whether aboveground vegetation chemistry and productivity, as detected from airborne sensors, predict soil properties, microbial processes and community composition. Imaging spectroscopy data were used to map aboveground biomass and green vegetation cover, functional traits and phylogenetic-functional community composition of vegetation. We examined the relationships between the image-derived variables and soil carbon and nitrogen concentration, microbial community composition, biomass and extracellular enzyme activity, and soil processes, including net nitrogen mineralization. In the BioDIV experiment—which has low overall diversity and productivity despite high variation in each—belowground processes were driven mainly by variation in the amount of organic matter inputs to soils. As a consequence, soil respiration, microbial biomass and enzyme activity, and fungal and bacterial composition and diversity were significantly predicted by remotely sensed vegetation cover and biomass. In contrast, at Wood River—where plant diversity and productivity were consistently higher—remotely sensed functional, chemical and phylogenetic composition of vegetation predicted belowground extracellular enzyme activity, microbial biomass, and net nitrogen mineralization rates. Aboveground biomass (or cover) did not predict these belowground attributes. The strong, contrasting associations between the quantity and chemistry of aboveground inputs with belowground soil processes and properties provide a basis for using imaging spectroscopy to understand belowground processes across productivity gradients in grassland systems. However, a mechanistic understanding of how above and belowground components interact among different ecosystems remains critical to extending these results broadly.
Response of boreal peatland ecosystems to global change: A remote sensing approach
(2017-08) McPartland, Mara
Global climate change is expected to result in anywhere from two to four degrees of warming, with consequences for terrestrial ecosystems. The rate of climate change is disproportionally greater at high latitudes, resulting in landscape-scale effects on the composition, structure, and function of arctic and boreal ecology. Remote sensing offers scientists the ability to track large-scale changes through the detection of biophysical processes occurring in terrestrial ecosystems. In this research, I measured the response of boreal peatland ecosystems to a suite of different climate-related drivers including increased temperature, elevated carbon dioxide levels, and hydrologic change. Working within large-scale ecosystem manipulation experiments, I used passive remote sensing to measure the response of two different types of boreal peatlands, a rich fen and an ombrotrophic bog, to simulated climate change. Chapter 1 describes my research on the use of hyperspectral remote sensing to examine changes in the composition and biodiversity of peatlands in response to long-term experimental manipulation. Chapter 2 details my findings on using simple remote sensing techniques to detect changes in peatland ecosystem productivity in response to warming, elevated carbon dioxide, and hydrologic change. Through this work, I demonstrate that remote sensing can be used to characterize the response of a range of different ecosystem properties to global change.
ROCR Isoprene Retrievals from the CrIS Satellite Sensor
(2022-02-28) Wells, Kelley C.; Millet, Dylan B.; dbm@umn.edu; Millet, Dylan B.; University of Minnesota Atmospheric Chemistry Group
This archive contains new atmospheric isoprene measurements from the CrIS satellite sensor using the ROCR retrieval algorithm. Data are archived in association with the manuscript below.
Using Single-Photon Lidar and Multispectral Imagery for Enhancing Forest Inventories
(2019-10) Allen, Benjamin
Modern forest management requires balancing multiple uses and management objectives, including timber production, wildlife habitat, and carbon sequestration. Forest inventories provide essential information for forest management decisions at a variety of spatial scales, including data about wood volume and the prevalence of various species. Traditional forest inventory systems rely primarily upon field data and design-based statistical estimators. These methods can provide unbiased estimates of inventory variables, albeit at a significant financial cost which limits the accuracy of the resulting data. Remote sensing technologies such as lidar and aerial photography have been used along with alternative statistical estimators to improve inventory accuracy and allow for spatially explicit maps of inventory data to be created. This research explored potential efficiency gains from the use of single-photon lidar and fall color aerial photography in a study area in northern Minnesota, USA. Remote sensing and field data combined in a model-assisted inferential framework were found to deliver relative efficiencies of approximately three for wood volume, with slightly lower values for basal area. Greater efficiency gains were found in coniferous-dominated forests than deciduous forests. The potential of these technologies to identify individual tree species and forest types was also examined. Classification between deciduous and coniferous-dominated forests provided overall classification accuracies of nearly 90% regardless of the classification algorithm used. By contrast, predictions of dominant species produced poor accuracy. Further research is needed to determine the economically optimal combination of remote sensing technologies for operational forest inventories.
Weather- and process-based models for the estimation of maize and soybean growth, development, and yield
(2020-01) Joshi, Vijaya
Field experiments in agricultural studies carried out at multiple sites and over several growing seasons are instrumental in improving crop management efforts. However, results from such experiments can have narrower applicability as results can vary depending on spatial and temporal variability in crop management practices, weather, and soil properties. In such context, crop models offer opportunities to overcome the shortcomings of field experiments conducted over limited periods and locations by simulating crop growth, development, and yield at various scenarios of weather and soil conditions. Three experiments were conducted to evaluate the application of crop models for maize and soybean production in the growing conditions of the US central Corn Belt. The first experiment evaluated the use and relative importance of readily available weather data to develop weather-based yield estimation models for maize and soybean. Total rainfall (Rain), average air temperature (Tavg), and the difference between maximum and minimum air temperature (Tdiff) at weekly, biweekly, and monthly time-scales from May to August were used to train multiple linear regression (MLR), general additive (GAM), and support vector machine (SVM) models to estimate county-level maize and soybean grain yields for Iowa, Illinois, Indiana, and Minnesota. For the total study area and at individual state level, SVM outperformed other models at all temporal levels for both maize and soybean. For maize, Tavg and Tdiff during July and August, and Rain during June and July were relatively more important whereas for soybean, Tavg in June and, Tdiff and Rain during August were more important weather variables to determine yield. The second experiment evaluated the simulation accuracy of the process-based CERES-Maize model. A study on four nitrogen (N) fertilizer rates for maize production was conducted during the growing seasons of 2016 and 2017 in southwest (Lamberton) and southern (Waseca) Minnesota. The model accurately simulated the dates of anthesis and maturity at both locations with a normalized root mean square error (nRMSE) of 1%. At Lamberton, final grain yield in both years was simulated within 16% nRMSE, but aboveground biomass was simulated with nRMSE as high as 30% and aboveground shoot N content and leaf area index (LAI) were simulated with nRMSEs as high as 38%. At Waseca, however, aboveground biomass over the growing season and final grain yield in both years were simulated with a 15% nRMSE, and aboveground shoot N content and LAI at both years were simulated with 21% nRMSE. Overall, the accuracy of the model was better with optimal growing conditions compared to no N fertilization. The third experiment compared the site-specific maize grain yield estimation accuracy of a stand-alone crop model, CERES-Maize, with a data-integration approach. In the integration approach, maize biomass estimated using satellite multispectral data at the five (V5) and ten (V10) leaf-collar stages were used to optimize the total soil nitrogen concentration (SLNI) and soil fertility factor (SLPF) in CERES-Maize. Without integration, maize yield was simulated with RMSE of 1264 kg ha-1. Optimization of SLNI improved yield simulations at both V5 and V10. However, better simulations were obtained from optimization at V10 as compared to V5. Optimization of SLPF together with SLNI did not further improve the yield simulations.

University Digital Conservancy

Browse by Subject

Browsing by Subject "remote sensing"