|
|
Regular Research Article 13 May 2013
Duke University, Durham, NC 27708, USA
Received: 09 Mar 2013 – Accepted: 15 Apr 2013 – Published: 13 May 2013 Abstract. Herein I discuss common errors in applying regression models and
wavelet filters used to analyze geophysical signals. I demonstrate that:
(1) multidecadal natural oscillations (e.g. the quasi 60 yr Multidecadal
Atlantic Oscillation (AMO), North Atlantic Oscillation (NAO) and Pacific
Decadal Oscillation (PDO)) need to be taken into account for properly
quantifying anomalous background accelerations in tide gauge records such as
in New York City; (2) uncertainties and multicollinearity among climate
forcing functions also prevent a proper evaluation of the solar contribution
to the 20th century global surface temperature warming using overloaded
linear regression models during the 1900–2000 period alone; (3) when
periodic wavelet filters, which require that a record is pre-processed with
a reflection methodology, are improperly applied to decompose non-stationary solar and
climatic time series, Gibbs boundary artifacts emerge yielding misleading
physical interpretations. By correcting these errors and using optimized
regression models that reduce multicollinearity artifacts, I found the
following results: (1) the relative sea level in New York City is not
accelerating in an alarming way, and may increase by about 350 ± 30 mm
from 2000 to 2100 instead of the previously projected values varying from
1130 ± 480 mm to 1550 ± 400 mm estimated using the methods proposed,
e.g., by Sallenger
Jr. et al. (2012) and Boon (2012), respectively; (2) the solar
activity increase during the 20th century contributed at least about 50%
of the 0.8 °C global warming observed during the 20th century instead
of only 7–10% (e.g.: IPCC, 2007; Benestad and Schmidt, 2009; Lean and Rind, 2009; Rohde et al., 2013). The first
result was obtained by using a quadratic polynomial function plus a 60 yr
harmonic to fit a required 110 yr-long sea level record. The second result
was obtained by using solar, volcano, greenhouse gases and aerosol
constructors to fit modern paleoclimatic temperature reconstructions
(e.g.: Moberg et al., 2005; Mann et al.,
2008; Christiansen and Ljungqvist, 2012) since the Medieval Warm Period,
which show a large millennial cycle that is well correlated to the millennial
solar cycle (e.g.: Kirkby, 2007; Scafetta and West, 2007; Scafetta,
2012c). These
findings stress the importance of natural oscillations and of the sun to
properly interpret climatic changes.
Citation: Scafetta, N.: Discussion on common errors in analyzing sea level accelerations, solar trends and global warming, Pattern Recogn. Phys., 1, 37-57, doi:10.5194/prp-1-37-2013, 2013.
|
|