Fig. 1. Schematic diagram showing data flow and analysis. Models were trained in the testing period (2006 - 2013) by three different types of input data sets as follows, Type 1: air temperature (Tair), relative humidity (RH), daily net radiation (Rd), precipitation (PPT), and evapotranspiration (ET) from eddy flux measurement, Type 2: air temperature (T), daily shortwave radiation (Rsd), and vapor pressure deficit (VPD) from MODIS, and Type 3: air temperature (T), daily shortwave radiation (Rsd), and vapor pressure deficit (VPD), and EVI from MODIS. Gross Primary Production (GPP) calculated based on multiple linear regression model (LM), support vector machine (SVM), random forest (RF), and artificial neural network (ANN) and evaluated with eddy covariance GPP in 2014 and 2015.
Fig. 2. Daily GPP prediction obtained with linear regression model (LM), Support Vector Machine (SVM), Random Forest (RF), and Artificial Neural Network (ANN) (black opened circle) and GPP obtained eddy covariance (EC) measurement (gray closed circle) using EC measurement datasets in 2014 and 2015.
Fig. 3. Comparisons of Eddy Covariance (EC) measurement GPP and modeled GPP from the trained models by EC measurement datasets in 2014 and 2015. LM = linear regression model, SVM = Support Vector Machine, RF = Random Forest, ANN = Artificial Neural Network.
Fig. 4. Daily GPP prediction obtained with linear regression model (LM), Support Vector Machine (SVM), Random Forest (RF), and Artificial Neural Network (ANN) (black opened circle) and GPP obtained eddy covariance (EC) measurement (gray closed circle) using MODIS datasets (without Enhanced Vegetation Index, EVI) in 2014 and 2015.
Fig. 5. Comparisons of Eddy Covariance (EC) measurement GPP and modeled GPP from the trained models by MODIS datasets (without Enhanced Vegetation Index, EVI) in 2014 and 2015. LM = linear regression model, SVM = Support Vector Machine, RF = Random Forest, ANN = Artificial Neural Network.
Fig. 6. Daily GPP prediction obtained with linear regression model (LM), Support Vector Machine (SVM), Random Forest (RF), and Artificial Neural Network (ANN) (black opened circle) and GPP obtained eddy covariance (EC) measurement (gray closed circle) using MODIS datasets (with EVI) in 2014 and 2015.
Fig. 7. Comparisons of Eddy Covariance (EC) measurement GPP and modeled GPP from the trained models by MODIS datasets (with EVI) in 2014 and 2015. LM = linear regression model, SVM = Support Vector Machine, RF = Random Forest, ANN = Artificial Neural Network.
Table 1. Summary of MODIS Land Products used in explanatory variables from 2006 to 2015
Table 2. Summary of MODIS Atmosphere Products used in explanatory variables from 2006 to 2015
Table 3. Table Summary of statistics for the different algorithms based on the three different input sets. R, RMSE, STD and MSE denote correlation coefficient, root mean square error, standard deviation and mean squared error, respectively
References
- Baldocchi, D., E. Falge, L. Gu, R. Olson, D. Hollinger, S. Running, P. Anthoni, C. Bernhofer, K. Davis, R. Evans, J. Fuentes, A. Goldstein, G. Katul, B. Law, X. Lee, Y. Malhi, T. Meyers, W. Munger, W. Oechel, K. T. Paw U, K. Pilegaard, H. P. Schmid, R. Valentini, S. Verma, T. Vesala, K. Wilson, and S. Wofsy, 2001: FLUXNET: a new tool to study the temporal and spatial variability of ecosystem-scale carbon dioxide, water vapor, and energy flux densities. American Meteorological Society 82(11), 2415-2434. https://doi.org/10.1175/1520-0477(2001)082<2415:FANTTS>2.3.CO;2
- Beck, P. S. A., C. Atzberger, K. A. Hogda, B. Johansen, and A. K. Skidmore, 2006: Improved monitoring of vegetation dynamics at very high latitudes: A new method using MODIS NDVI. Remote Sensing of Environment 100(3), 321-334. https://doi.org/10.1016/j.rse.2005.10.021
- Beer, C., M. Reichstein, E. Tomelleri, P. Ciais, M. Jung, N. Carvalhais, C. Rodenbeck, M. A. Arain, D. Baldocchi, G. B. Bonan, and A. Bondeau, 2010: Terrestrial gross carbon dioxide uptake: global distribution and covariation with climate. Science 329(5993), 834-838. https://doi.org/10.1126/science.1184984
- Bird, R. E., and R. L. Hulstrom, 1981: Simplified clear sky model for direct and diffuse insolation on horizontal surfaces (No. SERI/TR-642-761). Solar Energy Research Institute, Golden, CO (USA).
- Borbas, E. E., S. W. Seemann, A. Kern, L. Moy, J. Li, L. Gumley, and W. P. Menzel, 2011: MODIS atmospheric profile retrieval algorithm theoretical basis document. technical report, gfsc. nasa.
- Breiman, L., 1996: Bagging predictors. Machine learning 24(2), 123-140. https://doi.org/10.1023/A:1018054314350
- Breiman, L., 2001: Random Forests. Machine Learning 45(1), 5-32. https://doi.org/10.1023/A:1010933404324
- Cortes, C., and V. Vapnik, 1995: Support-Vector Networks. Machine Learning 20(3), 273-297. https://doi.org/10.1007/BF00994018
- Dimitriadou, E., K. Hornik, F. Leisch, D. Meyer, and A. Weingessel, 2008: Misc functions of the Department of Statistics (e1071), TU Wien. R package 1, 5-24.
- Dingman, S. L., 2008: Physical hydrology. Second Edition ed. Long Grove, Illinois, USA: Waveland Press, 636pp.
- Dou, X., Y. Yang, and J. Luo, 2018: Estimating forest carbon fluxes using machine learning techniques based on eddy covariance measurements. Sustainability 10, 203pp. https://doi.org/10.3390/su10010203
- Fritsch, S., F. Guenther, and M. F. Guenther, 2019: Package 'neuralnet'. Training of Neural Networks.
- Gray, A. N., and T. R. Whittier, 2014: Carbon stocks and changes on Pacific Northwest national forests and the role of disturbance, management, and growth. Forest ecology and management 328, 167-178. https://doi.org/10.1016/j.foreco.2014.05.015
-
Houborg, R. M., and H. Soegaard, 2004: Regional simulation of ecosystem
$CO_2$ and water vapor exchange for agricultural land using NOAA AVHRR and Terra MODIS satellite data. Application to Zealand, Denmark. Remote Sensing of Environment 93(1-2), 150-167. https://doi.org/10.1016/j.rse.2004.07.001 -
Ichii, K., M. Ueyama, M. Kondo, N. Saigusa, J. Kim, M. C. Alberto, J. Ardo, J., E.S. Euskirchen, M. Kang, T. Hirano, and J. Joiner, 2017: New data-driven estimation of terrestrial
$CO_2$ fluxes in Asia using a standardized database of eddy covariance measurements, remote sensing data, and support vector regression. Journal of Geophysical Research: Biogeosciences 122(4), 767-795. https://doi.org/10.1002/2016JG003640 - Jang, K., S. Kang, and S. Y. Hong, 2014: Comparisons of Collection 5 and 6 Aqua MODIS07_L2 air and dew temperature products with ground-based observation dataset. Korean Journal of Remote Sensing 30(5), 571-586. https://doi.org/10.7780/kjrs.2014.30.5.3
- Jang, K., S. Kang, J. S. Kimball, and S. Y. Hong, 2014b: Retrievals of all-weather daily air temperature using MODIS and AMSR-E data. Remote Sensing 6(9), 8387-8404. https://doi.org/10.3390/rs6098387
- Jang, K., M. Won, and S. Yoon, 2017: Evaluation of the satellite-based air temperature for all sky conditions using the Automated Mountain Meteorology Station (AMOS) records: Gangwon Province case study. Korean Journal of Agricultural and Forest Meteorology 19(1), 19-26. https://doi.org/10.5532/KJAFM.2017.19.1.19
- Kang, M., H. Kwon, J.-H. Cheon, and J. Kim, 2012: On estimating wet canopy evaporation from deciduous and coniferous forests in the Asian monsoon climate. Journal of Hydrometeorology 13(3), 950-965. https://doi.org/10.1175/JHM-D-11-07.1
- Kang, M., J. Kim, H.-S. Kim, B. M. Thakuri, and J.-H. Chun, 2014: On the nighttime correction of CO2 flux measured by eddy covariance over temperate forests in complex terrain. Korean Journal of Agricultural and Forest Meteorology 16(3), 233-245. https://doi.org/10.5532/KJAFM.2014.16.3.233
- Kang, M., J. Kim, B. M. Thakuri, J. Chun, and C. Cho, 2018: New gap-filling and partitioning technique for H2O eddy fluxes measured over forests. Biogeosciences 15(2), 631-647. https://doi.org/10.5194/bg-15-631-2018
- Kim, N. S., D. Han, J. Y. Cha, Y. S. Park, H. J. Cho, H. J. Kwon, Y.-C. Cho, S.-H. Oh, and C. S. Lee, 2015: A detection of novel habitats of Abies koreana by using species distribution models (SDMs) and its application for plant conservation. Journal of the Korea Society of Environmental Restoration Technology 18(6), 135-149. https://doi.org/10.13087/KOSERT.2015.18.6.135
- Kwon, H., S. Park, M. Kang, J. Yoo, R. Yuan, and J. Kim, 2007: Quality control and assurance of eddy covariance data at the two KoFlux sites. Korean Journal of Agricultural and Forest Meteorology 9(4), 260-267. https://doi.org/10.5532/KJAFM.2007.9.4.260
- Lantz, B., 2013: Machine learning with R. Packt Publishing Ltd., 125-135.
- Lee, J. H., S. Kang, K. C. Jang, J. H. Ko, and S. Y. Hong, 2011: The evaluation of meteorological inputs retrieved from MODIS for estimation of gross primary productivity in the US corn belt region. Korean Journal of Remote Sensing 27(4), 481-494. https://doi.org/10.7780/kjrs.2011.27.4.481
- Lee, B., W. Kang, C.-K. Kim, G. Kim, and C.-H. Lee, 2017: Estimating carbon uptake in forest and agricultural ecosystems of Korea and other countries using eddy covariance flux data. Journal of Environmental Impact Assessment 26(2), 127-139. https://doi.org/10.14249/eia.2017.26.2.127
- Lee, B., E. Kim, J. Lee, J.-M. Chung, and J.-H. Lim, 2018: Detecting phenology using MODIS vegetation indices and forest type map in South Korea. Korean Journal of Remote Sensing 34(2), 267-282. https://doi.org/10.7780/kjrs.2018.34.2.1.9
- Liaw, A., and M. Wiener, 2002: Classification and regression by randomForest. R news 2(3), 18-22.
- Masuoka, E., A. Fleig, R.E. Wolfe, and F. Patt, 1998: Key characteristics of MODIS data products. IEEE Transactions on Geoscience and Remote Sensing 36(4), 1313-1323. https://doi.org/10.1109/36.701081
- Oh, H. J., 2010: Landslide detection and landslide susceptibility mapping using aerial photos and artificial neural networks. Korean journal of remote sensing 26(1), 47-57.
- Recknagel, F., 2001: Applications of machine learning to ecological modelling. Ecological Modelling 146(1-3), 303-310. https://doi.org/10.1016/S0304-3800(01)00316-7
- Reeves, M. C., M. Zhao, and S. W. Running, 2005: Usefulness and limits of MODIS GPP for estimating wheat yield. International Journal of Remote Sensing 26(7), 1403-1421 https://doi.org/10.1080/01431160512331326567
- Running, S. W., D. D. Baldocchi, D. P. Turner, S. T. Gower, P. S. Bakwin, and K. A. Hibbard, 1999: A global terrestrial monitoring network integrating tower fluxes, flask sampling, ecosystem modeling and EOS satellite data. Remote Sensing of Environment 70(1), 108-127. https://doi.org/10.1016/S0034-4257(99)00061-9
- Running, S. W., P. E. Thornton, R. Nemani, and J. M. Glassy, 2000: Global terrestrial gross and net primary productivity from the Earth Observing System. In Methods in ecosystem science, 44-57. Springer, New York, NY (USA).
- Running, S. W., and M. Zhao, 2015: Daily GPP and annual NPP (MOD17A2/A3) products NASA Earth Observing System MODIS land algorithm. MOD17 User's Guide.
- Ryu, Y., S. Kang, S. K. Moon, and J. Kim, 2008: Evaluation of land surface radiation balance derived from moderate resolution imaging spectroradiometer (MODIS) over complex terrain and heterogeneous landscape on clear sky days. agricultural and forest meteorology 148(10), 1538-1552. https://doi.org/10.1016/j.agrformet.2008.05.008
- Schindler, D. E., and R. Hilborn, 2015: Prediction, precaution, and policy under global change. Science 347(6225), 953-954. http://doi.org/10.1126/science.1261824
- Sims, D. A., A. F. Rahman, V. D. Cordova, B. Z. El-Masri, D. D. Baldocchi, L. B. Flanagan, A. H. Goldstein, D. Y. Hollinger, L. Misson, R. K. Monson, and W. C. Oechel, 2006: On the use of MODIS EVI to assess gross primary productivity of North American ecosystems. Journal of Geophysical Research: Biogeosciences 111(G4).
- Smith, P., G. Lanigan, W. L. Kutsch, N. Buchmann, W. Eugster, M. Aubinet, E. Ceschia, P. Beziat, J. B. Yeluripati, B. Osborne, and E. J. Moors, 2010: Measurements necessary for assessing the net ecosystem carbon budget of croplands. Agriculture, ecosystems and environment 139(3), 302-315. https://doi.org/10.1016/j.agee.2010.04.004
- Smola, A. J., and B. Scholkopf, 2004: A tutorial on support vector regression. Statistics and computing 14(3), 199-222. https://doi.org/10.1023/B:STCO.0000035301.49549.88
- Tramontana, G., K. Ichii, G. Camps-Valls, E. Tomelleri, and D. Papale, 2015: Uncertainty analysis of gross primary production upscaling using Random Forests, remote sensing and eddy covariance data. Remote Sensing of Environment 168, 360-373. https://doi.org/10.1016/j.rse.2015.07.015
- Wang, X., Y. Yao, S. Zhao, K. Jia, X. Zhang, Y. Zhang, L. Zhang, and X. Chen, 2017: MODIS-based estimation of terrestrial latent heat flux over North America using three machine learning algorithms. Remote Sensing 9(12), 1326pp. https://doi.org/10.3390/rs9121326
- White, M. A., P. E. Thornton, S. W. Running, and R. R. Nemani, 2000: Parameterization and sensitivity analysis of the BIOME-BGC terrestrial ecosystem model: net primary production controls. Earth interactions 4(3), 1-85. https://doi.org/10.1175/1087-3562(2000)004<0001:ITEASI>2.3.CO;2
- Vidal, R., J. Bruna, R. Giryes, and S. Soatto, 2017: Mathematics of deep learning. arXiv preprint arXiv:1712.04741.
- Ye, H., R. J. Beamish, S. M. Glaser, S. C. Grant, C. H. Hsieh, L. J. Richards, J. T. Schnute, and G. Sugihara, 2015: Equation-free mechanistic ecosystem forecasting using empirical dynamic modeling. Proceedings of the National Academy of Sciences 112(13), E1569-E1576. https://doi.org/10.1073/pnas.1417063112