Clinton M. Rowe

Small-Scale Climate Maps: A Sensitivity Analysis of Some Common Assumptions Associated with Grid-Point Interpolation and Contouring

From Shepards (1968) local-search method, algorithms are developed for contouring on spherical surfaces and in Cartesian two-space. These algorithms are used to investigate errors on small-scale climate maps caused by the common practice of interpolating—from irregularly-spaced data points to regular-lattice nodes—and contouring in Cartesian two-space. Using mean annual air temperatures drawn from 100 irregularly-spaced weather stations, the annual air-temperature field over the western half of the northern hemisphere is estimated both on the sphere (assumed to be correct) and in Cartesian two-space. When these fields are mapped and compared, error magnitudes as large as 5° to 10° C appear in the air-temperature field approximated in Cartesian two-space.

Annual monsoon rains recorded by Jurassic dunes

Pangaea, the largest landmass in the Earths history, was nearly bisected by the Equator during the late Palaeozoic and early Mesozoic eras. Modelling experiments and stratigraphic studies have suggested that the supercontinent generated a monsoonal atmospheric circulation that led to extreme seasonality, but direct evidence for annual rainfall periodicity has been lacking. In the Mesozoic era, about 190 million years ago, thick deposits of wind-blown sand accumulated in dunes of a vast, low-latitude desert at Pangaeas western margin. These deposits are now situated in the southwestern USA. Here we analyse slump masses in the annual depositional cycles within these deposits, which have been described for some outcrops of the Navajo Sandstone. Twenty-four slumps, which were generated by heavy rainfall, appear within one interval representing 36 years of dune migration. We interpret the positions of 20 of these masses to indicate slumping during summer monsoon rains, with the other four having been the result of winter storms. The slumped lee faces of these Jurassic dunes therefore represent a prehistoric record of yearly rain events.

Long‐Lived Pluvial Episodes during Deposition of the Navajo Sandstone

The Navajo Sandstone of the American Southwest was deposited at approximately 190 Ma in a giant, subtropical dune field near the western margin of Pangea. From this unit, we report thick intervals of dune cross‐strata that were churned by insects and trampled by reptiles. Although dunes continued to migrate freely, the distribution of trace fossils shows that plant life in wet interdune areas sustained high levels of animal activity on the dunes for many thousands of years. We interpret this suite of structures as the record of a pluvial episode climatologically similar to the period of “greening” in the Sahara 4000–10,000 yr ago. A high percentage of the rainfall on the Navajo erg recharged the water table and led to the development of highly dilute, local groundwater flow systems that discharged into interdune areas.

Simulation of summer snowmelt on the Greenland ice sheet using a one‐dimensional model

A one-dimensional heat and mass balance model of a snowpack over frozen soil was modified for use in glacial environments. The model solves a set of governing equations for the energy and mass balances of the snow, subject to observed meteorological conditions at the upper boundary and the assumption of a steady state at the lower boundary. The initial state of the snowpack is defined by the temperature, density and grain size profiles at the beginning of the simulation period. The data used to test the model on the Greenland ice sheet are a subset of the meteorological and surface data collected during the 1990 summer field season by the Swiss Federal Institute of Technology (ETH) Greenland Expedition. The site was located near the equilibrium line elevation on the west slope of the ice sheet. The relatively large amount of snowmelt experienced at this site during the summer of 1990 provides a robust test of the snowmelt model. Both the simulated height and mass of the snowpack agree well with the observations. The evolution of profiles of temperature, density and liquid water content also conform to our expectations of the physical changes taking place in the snowpack during melt. Results from the present model are also compared to those from a similar model and differences between the models are discussed.

Inconsistencies Between Pangean Reconstructions and Basic Climate Controls

The supercontinent Pangea dominated our planet from the Permian into the Jurassic. Paleomagnetic reconstructions have been used to estimate the latitudinal position of Pangea during this 100-million-year period. Atmospheric circulation, recorded by eolian sandstones in the southwestern United States, shows a broad sweep of northeasterly winds over their northernmost extent, curving to become northwesterly in the south: This evidence is consistent with paleomagnetic reconstructions of the region straddling the equator in the Early Permian but is at odds with its northward movement to about 20°N by the Early Jurassic. At least one of the following scenarios must be true: The latitude based on paleomagnetism is incorrect; the interpretation of how winds shaped the dunes is mistaken; the basic climate controls in the Jurassic were different from those of today; or the paleogeographic reconstructions available are insufficient to adequately reproduce the wind fields responsible for dune formation.

Incorporating landscape heterogeneity in land surface albedo models

The importance of incorporating landscape heterogeneity into climate models has been recognized by a number of researchers. However, attempts to relax the assumption of uniform, homogeneous and complete vegetation cover have generally been limited to modeling each component of the landscape (e.g., trees, grasses, shrubs) as a homogeneous unit and then combining the results for each type, weighted by the percentage of ground area it covers (i.e., an area-weighted average is computed). The purpose of this paper is to demonstrate that substantial errors in the representation of vegetation-atmosphere interactions can arise from the assumption of landscape homogeneity. Even when weighted averages are employed as a first attempt to consider landscape heterogeneity, the underlying assumption of homogeneity of the vegetation patches can lean to significant errors. To demonstrate an example of how this can occur, vegetation canopy albedo derived from weighted averages is compared to that computed with explicit consideration of canopy heterogeneity.

An Observational Analysis and Evaluation of Land Surface Model Accuracy in the Nebraska Sand Hills

In this study, the influence of subsurface water on the energy budget components of three locations with heterogeneous land surfaces in the Nebraska Sand Hills are examined through observations and use of the Noah land surface model (LSM). Observations of the four primary components of the surface energy budget are compared for a wet interdunal meadow valley, a dry interdunal valley, and a dunal upland location. With similar atmospheric forcing at each site, it was determined that differences in the partitioning of the mean diurnal net radiation (Rnet) existed among the three locations due to the influence of varied soil moisture and vegetation through the year. At the wet valley, observations indicated that almost 65% of the mean daily peak Rnet was used for latent heating, due to the relatively higher soil moisture content resulting from an annual upward gradient of subsurface water and denser vegetation. In sharp contrast, the dunal upland site yielded only 21% of the mean daily peak Rnet going to latent heating, and a greater mean diurnal soil heat flux with typically drier soils and sparser vegetation than at the wet valley. The dry valley partition of the peak Rnet fell between the wet valley and dunal upland site, with approximately 50% going to sensible heating and 50% toward latent heating. In addition to the observational analysis, an uncoupled land surface model was forced with the observations from each site to simulate the energy budgets, with no tuning of the model’s fundamental equations and with little adjustment of the model parameters to improve results. While the model was able to reasonably simulate the mean diurnal and annual energy budget components at all locations, in most instances with root-mean-square errors within 20%–25% of the observed values, the lack of explicit treatment of subsurface water within the model limited predictability, particularly at the wet valley site. For instance, only 25% of the peak mean diurnal Rnet went toward latent heating in the model simulation of the wet valley, compared to 65% as estimated by observations. Model evaluation statistics are presented to document the land surface model’s ability to capture the annual and mean diurnal variations in the surface energy budget terms at the dry valley and dunal upland sites, but the absence of subsurface water results in large errors in the wet valley simulation. From these results, a case is made for the future inclusion of the explicit treatment of subsurface water within the Noah LSM to better approximate the prediction of the surface energy budget in such environments.

MODELING LAND-SURFACE ALBEDOS FROM VEGETATION CANOPY ARCHITECTURE

Over the past decade, the climatic impact of changes in terrestrial albedo has been studied using numerous climate models, ranging from simple, one-dimensional energy balance climate models to the more sophisticated, three-dimensional general circulation models of the atmosphere. In most of these, however, the land-surface albedos have been prescribed both spatially and temporally from albedo observations. To overcome the limitations of using prescribed land-surface albedos, a model of radiation transfer in plant canopies was used to predict vegetation albedos. Because of the models reliance on the physical properties of the land-surface cover, it is able to account explicitly for albedo variations caused by factors both internal to and external from the vegetation canopy. The model is described and the results of simulations for three representative canopy types are discussed. The dependence of albedo on irradiance distribution predicted by the model agrees well with established theory. [Key words: albe...

A comparison of microwave radiometric data and modeled snowpack conditions for Dye 2, Greenland

SummaryMeteorological observations were recorded at Dye 2, Greenland during the summer of 1993 as part of a research program to identify interannual variations in melt occurrence on the Greenland ice sheet from satellite microwave data. The meteorological observations were used to drive and energy-balance model of the snowpack during 21 June to 13 July 1993. Time series of the meteorological observations and various model outputs were compared to a concurrent time series of Special Sensor Microwave/Imager (SSM/I) data for scan cells centered within 25 km of Dye 2. The satellite microwave observations clearly show an increase in snowpack emissivity at the same time that the model indicates liquid water forming in the snow. Diurnal melt-freeze cycles that occurred during mid June to early July resulted in an increase in the 37 GHz brightness temperature as great as 60K from the dry, refrozen snow in the morning to the wet snow of some afternoons. The effects of fresh snowfall, which tend to increase the brightness temperature, and of snow growth from melt-freeze metamorphism, which tend to decrease the brightness temperature, are also apparent in the microwave observations. The results of this work demonstrate the influence of daily weather variations on the microwave emissivity in the ice sheets percolation zone and the usefulness of swath data to diagnose the diurnal cycle of melt.

A High-Resolution Modeling Strategy to Assess Impacts of Climate Change for Mesoamerica and the Caribbean

Mesoamerica and the Caribbean are low-latitude regions at risk for the effects of climate change. Global climate models provide large-scale assessment of climate drivers, but, at a horizontal resolution of 100 km, cannot resolve the effects of topography and land use as they impact the local temperature and precipitation that are keys to climate impacts. We developed a robust dynamical downscaling strategy that used the WRF regional climate model to downscale at 4 - 12 km resolution GCM results. Model verification demonstrates the need for such resolution of topography in order to properly simulate temperatures. Precipitation is more difficult to evaluate, being highly variable in time and space. Overall, a 36 km resolution is inadequate; 12 km appears reasonable, especially in regions of low topography, but the 4 km resolution provides the best match with observations. This represents a tradeoff between model resolution and the computational effort needed to make simulations. A key goal is to provide climate change specialists in each country with the information they need to evaluate possible future climate change impacts.