Skip to main content


Showing posts from June, 2016

Numerical simulation: unit system

Unit is always one of the most important factors in science. Most of the time, a number without unit is useless.
For numerical simulation, a good design of unit system can be critical.
Recently, I have written a short script/program to prepare climate data for my groundwater/surface water hydrology simulation. I downloaded some data from the National Climatic Data Center(NCDC) separately, and I did NOT notice the changes in unit system. In the end, I have to do everything over again.

So the question is: what kind of unit system should we use to improve our efficiency, and how do we actually implement it?

First of all, we need to identify the unit system used in the data obtained from third party. This is usually done through reading the meta data in the documentation. Never directly use the data without reading the documentation!
A lot of time we don't usually have choices of the unit system provided. In this case, the best practice might be stick with the original unit system.


Groundwater hydrology modeling: Why it is so difficult to simulate the groundwater dynamics in mountain area

First, in mountain area, there is generally less groundwater observations, which makes it's difficult to define the boundary condition. Without observations, the initial head can only be assumed or estimated using a steady state simulation.
Second, groundwater recharge in mountain area is very low. This is because some mountain surface has no soil or a shallow soil layer. And the permeability of rock fractures is very low. As a result, infiltration from the surface/subsurface to groundwater systems is extremely low.
Third, the hydraulic priorities of the fractured and unfractured rock material are also very low, which means the groundwater system is very sensitive.
Beside, topography effects from mountain area usually require high spatial resolution grid cell in the numerical simulation.

Ecosystem modeling: challenges in water-carbon cycling in cold region numerical simulation

Numerical simulation of water and carbon cycling in cold region usually encounter some unique challenges which must be taken into consideration in climate change study.

First, the presence of snow, glacier and permafrost has direct influence on the water cycling. Snow hydrology has been long studied and I try to avoid discussing it here. Glacier dynamics have also drawn a lot of attentions in recent decades due to the warming climate.
Permafrost also plays an important role, which has not yet been fully understood and investigated. Permafrost and its upper soil layer (Active Layer) have seasonal dynamics in water/ice content and temperature. As temperature continues to increase, the thawing permafrost also releases the ice to the surface layer, which means the active layer thickness is expected to increase in the near future.
Other than that, unlike that in temperate regions, permafrost acts as a confining layer between surface water and groundwater, which means the groundwater and su…

Ecosystem modeling: a question of chicken or eggs?

Ecosystem modeling generally include three conceptual components: inputs, algorithm and outputs.
For example, we usually need precipitation to estimate surface runoff.
On the other hand, sometimes some outputs are also considered as inputs for other models. For example, vegetation dynamics also change the surface albedo and therefore the incoming radiation.
In another word, the feedback among different processes are often too complex that we generally have to ignore some feedback processes.

The reason is that our computer simulation MUST have a starting point and a sequence of algorithms in order to run the simulation.
For some well-designed models, the differences between different starting points are not significant when using the numerical approach. However, it is a best practice to put all the algorithms in the right orders.

For example, in surface hydrology, the generally order of the water flow is like this: precipitation->interception->snow->infiltration->overland r…