Humans and Hydrology at High Latitudes (H 3 L) Richard B. Lammers Water Systems Analysis Group, UNH Dan WhiteUniversity of Alaska, Fairbanks Lawrence C. Hamilton Department of Sociology, UNH Lilian Alessa University of Alaska, Anchorage Alexander I. Shiklomanov Water Systems Analysis Group, UNH Charles J. Vorosmarty Water Systems Analysis Group, UNH Rasmus O. Rasmussen University of Roskilde, Denmark Igor A. Shiklomanov Director, State Hydrological Institute, St. Petersburg, Russia Cynthia M. Duncan University of New Hampshire Sponsored by NSF - Synthesis of Arctic System Science - OPP 0531148
33
Embed
Humans and Hydrology at High Latitudes (H 3 L) Richard B. Lammers Water Systems Analysis Group, UNH Dan WhiteUniversity of Alaska, Fairbanks Lawrence C.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Humans and Hydrology at High Latitudes (H3L)
Richard B. Lammers Water Systems Analysis Group, UNHDan White University of Alaska, FairbanksLawrence C. Hamilton Department of Sociology, UNHLilian Alessa University of Alaska, AnchorageAlexander I. Shiklomanov Water Systems Analysis Group, UNHCharles J. Vorosmarty Water Systems Analysis Group, UNHRasmus O. Rasmussen University of Roskilde, DenmarkIgor A. Shiklomanov Director, State Hydrological Institute, St. Petersburg, RussiaCynthia M. Duncan University of New Hampshire
Sponsored by NSF - Synthesis of Arctic System Science - OPP 0531148
Major Challenges
(1) Integrating biogeophysical and human dimension data sets at the pan-arctic scale
(2) Ranking the major forcings on the water system
(3) Drawing conclusions on current system state and future trajectories at scales relevant to human activities (i.e. local and intermediate)
Hypotheses
H1: At local scales, direct human impacts (such as changes in land use and land cover, or water management) most often drive hydrologic change. At larger regional and pan-arctic scales, climate change drives hydrologic change.
H2: Climate change affects human activities in the pan-Arctic in ways that will, in turn, impact both the humans and hydrology of the pan-arctic system.
H3: A more temperate pan-arctic basin will lead to conditions under which direct human impacts on the hydrologic cycle will intensify and expand in scale.
GOAL 1
Retrospective: To analyze the major forces and trajectories shaping the pan-arctic water system and to understand their interactions with humans.
GOAL 2
Contemporary: To advance our knowledge of relationships linking broad scales of change to local societal impacts.
GOAL 3
Future: To forecast the range of potential future statistics of the pan-arctic hydrosphere, societal impacts, and response at multiple scales.
The proposed work involves a 200 year time range made up of the historical (1900-2000),
contemporary (2000), and prognostic (2000-2100) periods. These periods can be viewed from the
perspective of a variety of scales; large (continental), regional, and local (human) scales.
Aleutians East Borough Aleutians West Census Area Bethel Census Area Bristol Bay Borough
Dillingham Census Area Lake and Peninsula Borough Nome Census Area North Slope Borough
Northwest Arctic Borough Wade Hampton Census AreaWrangell-Petersburg Census AreaYukon-Koyukuk Census Area
Population in 1000s
Observed & modeled population in 12 Alaska regions, 1969–2003
8 4 0 4 8 8 4 0 4 8
8 4 0 4 8 8 4 0 4 8 8 4 0 4 8 8 4 0 4 8
AK ALE ALW ANC BET BRB
DEN DIL FBX HAI JNU KEN
KET KOD LP MAT NOM NSB
NWA POW SHA SIT SOF VAL
WAD WRP YAK YUK
Multilevel models, also called mixed models, can include both “fixed effects,” which at the top level are analogous to coefficients of one-level models, and also “random effects,” which vary within levels. We can have multiple nested levels of random effects. The red curves in Figure 1 depict a simple two-level model of the form
pop i j = β0 + β1 year i j + β2 year 2i j + ζ0 j + ζ1 j year i j + ζ2 j year 2i j + ε i j
where pop i j is the population in year i for region j. The β (beta) parameters describe growth trends for all 12 regions considered together. ζ j (zeta) parameters represent effects that are unique for each region. Estimates of β’s and the standard deviations of ζ’s (all of which differ significantly from zero) are summarized in the following table.
. xtmixed pop year0 year2 || fips: year0 year2Performing EM optimization: Performing gradient-based optimization: Iteration 0: log restricted-likelihood = -2881.5464 Iteration 1: log restricted-likelihood = -2881.5464 Computing standard errors:
Mixed-effects REML regression Number of obs = 376Group variable: fips Number of groups = 12