Simulation of the growth ring density of Larix olgensis plantation wood with the ARMA model

Abstract

Because growth ring data have temporal features, time series analysis can be used to simulate and reveal changes in the life of a tree and contribute to plantation management. In this study, the autoregressive (AR) and moving average modeling method was used to simulate the time series for growth ring density in a larch plantation with different initial planting densities. We adopted the Box–Jenkins method for the modeling, which was initially based on an intuitive analysis of sequence graphs followed by the augmented Dickey–Fuller stationarity test. The order p and q of the ARMA (p, q) model was determined based on the autocorrelation and partial correlation coefficient figure truncated on the respective order. Through the residual judgment, the model AR (2) was only fitted to the larch growth ring density series for the plantation with the 1.5 × 2.0 m² initial planting density. Because the residuals series for the other three series was not shown as a white noise sequence, the modeling was rerun. Larch wood from the initial planting density of 2.0 × 2.0 m² was modeled by ARMA (2, 1), and ARMA ((1, 5), 3) fitted to the 2.5 × 2.5 m² initial planting density, and the 3.0 × 3.0 m² was modeled by AR (1, 2, 5). Although the ARMA modeling can simulate the change in growth ring density, data for the different growth ring time series were described by different models. Thus, time series modeling can be suitable for growth ring data analysis, revealing the time domain and frequency domain of growth ring data.

Keywords

Growth ring density / Larix olgensis plantation wood / ARMA / Modeling / Time series analysis

References

[1] Bowman DM, Brienen RJ, Gloor E, Phillips OL, Prior LD. Detecting trends in tree growth?. Trends Plant Sci, 2013, 18: 11-17.

[2] Box GEP, Jenkins GM. Time series analysis: forecasting and control, 1976 Revised Pasadena: Holden-day 161 215

[3] Box GEP, Pierce DA. Distribution of residual autocorrelations in autoregressive-integrated moving average time series models. J Am Stat Assoc, 1970, 65: 1509-1526.

[4] Briffa KR, Schweingruber FH, Jones PD. Trees tell of past climates: but are they speaking less clearly today?. Philos Trans R Soc Lond B Biol Sci, 1998, 353: 65-73.

[5] Dickey DA, Fuller WA. Distribution of the estimators for autoregressive time series with a unit root. J Am Stat Assoc, 1979, 74(366): 427-431.

[6] Esper J, Frank D, Buntgen U, Verstege A, Hantemirov R, Kirdyanov AV. Trends and uncertainties in Siberian indicators of 20th century warming. Glob Change Biol, 2010, 16: 386-398.

[7] Fritts HC. Tree rings and climate, 2001, Caldwell: Blackburn Press 511 529

[8] Hughes MK, Kelly PM, Pilcher JR. Climate from tree rings, 1982, New York: Cambridge University Press 217 219

[9] Jaakkola T, Mäkinen H, Saranpaä P. Wood density of Norway spruce: responses to timing and intensity of first commercial thinning and fertilization. For Ecol Manag, 2006, 237: 513-521.

[10] Jiang SL, Du LC. Height and diameter growth modeling of Dahurian Larch based on nonlinear mixed model in northeastern China. For Res, 2012, 25(1): 11-16.

[11] Kwiatkowski D, Phillips PCB, Schimit P, Shin Y. Testing the null hypothesis of stationary against the alternative of a unit root. J Econ, 1992, 54: 159-178.

[12] Liu CQ, Qi YF. The study of time series analysis and prediction model of larch trees. Anhui Agric Sci Bull, 2014, 20(16): 81-82.

[13] Ljung GM, Box GEP. On a measure of a lack of fit in time series models. Biometrika, 1978, 65(2): 297-303.

[14] MacKinnon JG. Numerical distribution functions for unit root and cointegration tests. J Appl Econ, 1996, 11: 601-618.

[15] Ng S, Perron P. Lag length selection and the construction of matrix estimation. Rev Econ Stud, 2001, 61: 631-653.

[16] Oliveira JM, Roig FA, Pillar VD. Climatic signals in tree-rings of Araucaria angustifolia in the southern Brazilian highlands. Austral Ecol, 2010, 35: 134-147.

[17] Pandit SM, Wu SM. Modeling and analysis of closed-Loop systems from operating data. Technometrics, 1977, 19: 477-485.

[18] Peters RL, Groenendijk P, Vlam M, Zuidema PA. Detecting long-term growth trends using tree rings: a critical evaluation of methods. Glob Change Biol, 2015, 21: 2040-2054.

[19] Phillips PB, Perron P. Testing for a unit root in time series regression. Biometrika, 1988, 75(2): 335-346.

[20] Rozendaal DMA, Brienen RJW, Soliz-Gamboa CC, Zuidema PA. Tropical tree rings reveal preferential survival of fast-growing juveniles and increased juvenile growth rates over time. New Phytol, 2010, 185: 759-769.

[21] Shi P, Körner C, Hoch G. A test of the growth-limitation theory for alpine tree line formation in evergreen and deciduous taxa of the eastern Himalayas. Funct Ecol, 2008, 22: 213-220.

[22] Sleen P, Groenendijk P, Vlam M. No growth stimulation of tropical trees by 150 years of CO₂ fertilization but water-use efficiency increased. Nat Geosci, 2015, 8: 24-28.

[23] Swaef TD, Schepper VD, Vandegehuchte MW, Steppe K. Stem diameter variations as a versatile research tool in ecophysiology. Tree Physiol, 2015, 35: 1047-1061.

[24] Villalba R, Lara A, Masiokas MH. Unusual southern hemisphere tree growth patterns induced by changes in the Southern Annular Mode. Nat Geosci, 2012, 5: 793-798.

[25] Wang JM, Zhu TL, Bi WQ. The time series model of the wood growth ring property variation pattern. J Northeast For Univ, 1998, 26(2): 45-48.

[26] Wu XD. Tree rings and climate change, 1990, Beijing: China Meteorological Press 316 328