Reconciling precipitation with runoff : Observed hydrological change in the midlatitudes
(2015) In Journal of Hydrometeorology 16(6). p.2403-2420- Abstract
Century-long observed gridded land precipitation datasets are a cornerstone of hydrometeorological research. But recent work has suggested that observed Northern Hemisphere midlatitude (NHML) land mean precipitation does not show evidence of an expected negative response to mid-twentieth-century aerosol forcing. Utilizing observed river discharges, the observed runoff is calculated and compared with observed land precipitation. The results show a near-zero twentieth-century trend in observed NHML land mean runoff, in contrast to the significant positive trend in observed NHML land mean precipitation. However, precipitation and runoff share common interannual and decadal variability. An obvious split, or breakpoint, is found in the NHML... (More)
Century-long observed gridded land precipitation datasets are a cornerstone of hydrometeorological research. But recent work has suggested that observed Northern Hemisphere midlatitude (NHML) land mean precipitation does not show evidence of an expected negative response to mid-twentieth-century aerosol forcing. Utilizing observed river discharges, the observed runoff is calculated and compared with observed land precipitation. The results show a near-zero twentieth-century trend in observed NHML land mean runoff, in contrast to the significant positive trend in observed NHML land mean precipitation. However, precipitation and runoff share common interannual and decadal variability. An obvious split, or breakpoint, is found in the NHML land mean runoff-precipitation relationship in the 1930s. Using runoff simulated by six land surface models (LSMs), which are driven by the observed precipitation dataset, such breakpoints are absent. These findings support previous hypotheses that inhomogeneities exist in the early-twentieth-century NHML land mean precipitation record. Adjusting the observed precipitation record according to the observed runoff record largely accounts for the departure of the observed precipitation response from that predicted given the real-world aerosol forcing estimate, more than halving the discrepancy from about 6 to around 2 W m-2. Consideration of complementary observed runoff adds support to the suggestion that NHML-wide early-twentieth-century precipitation observations are unsuitable for climate change studies. The agreement between precipitation and runoff over Europe, however, is excellent, supporting the use of whole-twentieth-century observed precipitation datasets here.
(Less)
- author
- publishing date
- 2015
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- Atm/Ocean Structure/ Phenomena, Changepoint analysis, Climate variability, Geographic location/entity, Land surface, Land surface model, Mathematical and statistical techniques, Models and modeling, Precipitation, Runoff, Variability
- in
- Journal of Hydrometeorology
- volume
- 16
- issue
- 6
- pages
- 18 pages
- publisher
- American Meteorological Society
- external identifiers
-
- scopus:84950131486
- ISSN
- 1525-755X
- DOI
- 10.1175/JHM-D-15-0055.1
- language
- English
- LU publication?
- no
- id
- 55f985ed-19d4-4eef-8463-ac2730709d70
- date added to LUP
- 2020-11-19 23:18:57
- date last changed
- 2022-02-01 17:53:38
@article{55f985ed-19d4-4eef-8463-ac2730709d70, abstract = {{<p>Century-long observed gridded land precipitation datasets are a cornerstone of hydrometeorological research. But recent work has suggested that observed Northern Hemisphere midlatitude (NHML) land mean precipitation does not show evidence of an expected negative response to mid-twentieth-century aerosol forcing. Utilizing observed river discharges, the observed runoff is calculated and compared with observed land precipitation. The results show a near-zero twentieth-century trend in observed NHML land mean runoff, in contrast to the significant positive trend in observed NHML land mean precipitation. However, precipitation and runoff share common interannual and decadal variability. An obvious split, or breakpoint, is found in the NHML land mean runoff-precipitation relationship in the 1930s. Using runoff simulated by six land surface models (LSMs), which are driven by the observed precipitation dataset, such breakpoints are absent. These findings support previous hypotheses that inhomogeneities exist in the early-twentieth-century NHML land mean precipitation record. Adjusting the observed precipitation record according to the observed runoff record largely accounts for the departure of the observed precipitation response from that predicted given the real-world aerosol forcing estimate, more than halving the discrepancy from about 6 to around 2 W m<sup>-2</sup>. Consideration of complementary observed runoff adds support to the suggestion that NHML-wide early-twentieth-century precipitation observations are unsuitable for climate change studies. The agreement between precipitation and runoff over Europe, however, is excellent, supporting the use of whole-twentieth-century observed precipitation datasets here.</p>}}, author = {{Osborne, Joe M. and Lambert, F. Hugo and Groenendijk, Margriet and Harper, Anna B. and Koven, Charles D. and Poulter, Benjamin and Pugh, Thomas A.M. and Sitch, Stephen and Stocker, Benjamin D. and Wiltshire, Andy and Zaehle, Sönke}}, issn = {{1525-755X}}, keywords = {{Atm/Ocean Structure/ Phenomena; Changepoint analysis; Climate variability; Geographic location/entity; Land surface; Land surface model; Mathematical and statistical techniques; Models and modeling; Precipitation; Runoff; Variability}}, language = {{eng}}, number = {{6}}, pages = {{2403--2420}}, publisher = {{American Meteorological Society}}, series = {{Journal of Hydrometeorology}}, title = {{Reconciling precipitation with runoff : Observed hydrological change in the midlatitudes}}, url = {{http://dx.doi.org/10.1175/JHM-D-15-0055.1}}, doi = {{10.1175/JHM-D-15-0055.1}}, volume = {{16}}, year = {{2015}}, }