December and January DLI in Everglades City, Florida

08 February 2016

I've been reading about the rains and clouds in South Florida and how extraordinary the past couple months have been. I saw these charts from Travis Shaddox, and I wondered what the light would be in photosynthetic units.

I downloaded monthly summary data since March 2007 for Everglades City from the NOAA. I use these data because they include global solar radiation, and I converted from energy units of MJ/m2 to photosynthetic units of mol/m2 using the 2.04 conversion factor of Meek et al.

This shows the average daily light integral (DLI) each month. One can see the seasonal changes, and one can also see that December 2015 had the lowest DLI of any December and that January 2016 had the lowest DLI of any January. I plotted all the data I could get, which is since 2007; I don't know what the values would have been before that. In the past decade, though, these were the lowest.

Looking just at December and January year by year, January 2016 really stands out for having a low DLI. Blue triangles are December DLIs and red circles are January DLIs; the vertical dashed lines (blue for December, red for January) show the averages prior to Dec 2015 and Jan 2016.

In a normal year at Everglades City, January would have more photosynthetic light than December. For seven out of the past eight years, the month of December had a lower DLI than January.  Only 2014 had a lower DLI in January than in December. But January 2016 is a big outlier; not only does January 2016 have the lowest DLI of any of the previous Januaries, but it also has a lower DLI than any of the previous Decembers.

GP and GDD: are they comparable?

19 January 2016

Someone asked me at the Northern Green Expo if the temperature-based growth potential (GP) and temperature-based growing degree days (GDD) are comparable. They sort of are, with a couple of exceptions. Comparable, yeah, kind of. But they are not interchangeable.

I downloaded the weather data for every day in 2015 from the international airports at London (Heathrow), Minneapolis, Sydney, and Tokyo (Haneda). Then I calculated the GP, and the GDD, and I made the charts shown in this post. The script to download the data and produce the charts is here . I’ll try to explain this, but I think it is easiest to see how GP and GDD are similar, and how they differ, by making some comparisons yourself.

First, here are the mean daily temperatures in 2015. The points are daily mean temperatures for each day of the year, and the lines are a moving average. Sydney and Tokyo are both hot in the summer, Minneapolis is coldest in the winter and hotter than London in the summer, and London is coolest in the summer but has winter temperatures close to those of Tokyo.

Those cities have fairly diverse temperature ranges and variation in temperature from winter to summer. One expects a different growing environment in each. The GP3 is a value with a minimum of 0 and a maximum of 1, showing the expected limitation (or potential) of temperature on growth.

What do we see there? It’s a bit different than the temperature. Looking at the moving average for each city, we see Tokyo has a big drop in mid-summer because it is too hot, and Sydney has a substantial drop too, and Minneapolis has a slight drop in GP during the hottest summer temperatures, and London has peak GP in mid-summer because the average temperature rarely exceeds the optimum growth temperatures.

In the winter, the GP3 drops to almost 0 at Minneapolis, London, and Tokyo, but at Sydney it drops just below 0.5 in mid-winter, indicating that C3 grasses should still be able to grow, albeit slowly.

That was GP through the year. Now we can look at GDD0 . That is, for each day with an average temperature above 0°C (32°F), I take that temperature and call that the GDD. In this case, I would be using a base temperature of 0°C. This is the basis for the growing degree day model of Kreuser for the reapplication of plant growth regulators.

That’s not exactly like the GP plot above. It is like zooming in on the temperature chart, but only showing the portion of the chart with temperatures above 0°C. Compared to the GP chart, one notices that with GDD 0 there is no drop in mid-summer when it is too hot, and the GDD0 does not drop all the way to 0 in winter at Tokyo and London.

So far it seems the GP and GDD are sort of the same, and sort of different. Both are based on temperature. But GDD is a measure of heat accumulation. GP is generating a value with a minimum of 0 when temperatures are far from an optimum for photosynthesis, and then generating a value that gets closer to 1 as the temperatures get progressively closer to 1.

There are various ways to calculate the heat accumulation through growing degree days. The GDD10 only counts the degrees on those days when the average temperature is above 10°C (50°F). This is GDD with a base temperature of 10°C. That makes sense for some things, and this chart of GDD10 is similar to the GDD0 chart in that it is as if the temperature chart were cropped to omit all values less than 10°C.

That’s what GDD10 is showing. Now we are looking only at the days in the year when the average temperature was above 10°C, and we can see how much heat accumulation there would be each day.

The big difference between GP and GDD is evident, because Sydney and Tokyo are peaking in GDD when temperatures are at their hottest, but GP would produce a value less than 1 when GDD was highest in those places, because the temperatures are considered too hot for optimum growth of C3 grass.

GDD is heat accumulation. GP is optimum growth temperature accumulation. Let’s look at the accumulation explicitly, by adding together the GP for every day of the year in order to get these lines showing the cumulative sum of GP in 2015.

Sydney with the year-round growth, although with a dip in winter and also a dip in summer, has the highest sum of GP. Then Tokyo, and Minneapolis and London are similar.

We can do the same type of chart for GDD0 .

Sydney still has the highest sum, then Tokyo, but there is less of a distance between these two cities than with GP, because GDD is using all of Tokyo’s hot summer days, but GP in Tokyo drops when it is hot. London and Minneapolis are similar again, but notice that Minneapolis accumulates almost all its GDD0 from April to October, while the milder winter in London allows the GDD0 to accumulate slowly year-round.

The cumulative sum of GDD50 is just a little different.

Now Tokyo catches and exceeds Sydney in the northern hemisphere autumn, but Sydney catches up quickly as summer approaches. And when counting the heat accumulation now only above 10°C, Minneapolis now has a lot more of that than does London.

The GP and GDD with various base temperatures (0 and 10°C are two of the standard ones) can be used for different things. GDD is good for things that are heat dependent. Growth regulators, insects, diseases, weeds – certainly the growth of certain plants in the range of temperatures from the minimum temperature required for growth up to the optimum temperature for growth. The GP is formulated in a different way, where it decreases when it is too cold or too hot for optimum growth.

We can look at that a little more closely. Now let’s just look at 2 cities to reduce the overlap: London and Minneapolis. Here is the GDD0 for every day of 2015.

That’s a linear increase in GDD0 for every increase in temperature above 0°C. If we would plot GDD10 , there would also be a linear increase with temperature, but the line would start going up at a mean daily temperature of 10 rather than at 0.

The GP for that same range of temperatures at London and Minneapolis looks completely different.

That’s because the GP has a minimum of 1 and a maximum of 0, and the value is dependent on how close the temperature is to the optimum temperatures for photosynthesis.

Now to get back to the original question, after all those examples, are GDD and GP comparable? For them to be comparable, there would have to be a linear relationship (or almost linear relationship) between the accumulated GP and the accumulated GDD through the year.

Here I’ve plotted just that; the cumulative sum of GP for 2015 is on the x-axis, and the cumulative sum of GDD0 is on the y-axis.

Well, that is sort of linear, but has a few weird curves or shifts. London’s GDD goes up in winter when the GP is still low, and Tokyo’s GDD goes way up in mid-summer when the accumulated GP is increasing slowly. And the line for Sydney looks pretty straight by comparison, but we can take a closer look at that by checking GP vs GDD10 .

Now we are looking at how GP accumulates through the year, and comparing that to how GDD10 accumulates. At some times of the year it is linear, but as temperatures get low or high, the slope of that line changes.

If you would make these calculations for data at your location, I think you would see the same thing and would see how GP and GDD are similar and how they are different.

"How do you calculate how much water is needed for a given area?"

18 January 2016

I received this inquiry last week:

"When looking at water quantity for a new golf course, you have to determine how much water is needed obviously so what I want to know is.

1) How do you calculate how much water is needed for a given area (the whole golf course?)

2)  I know you have to look at the driest year data and base it on that but I understand you have to measure how much a grass plant and soil will lose via Evapotranspiration so you know what you have to replace so what methods do you use to find this out?"

I suggest calculating a water budget for the location using the method described by Gelernter et al. with this supplement providing the details for the calculations.

A summary of photosynthetically active radiation for 1 year at Tokyo

01 January 2016

This chart (download it here) shows the average photosynthetic photon flux density (PPFD) for each hour of 2015 at Tokyo. The daily light integral (DLI) is the number written in black at the top right corner of each facet in this chart.

One can get some idea of how the DLI changes seasonally and with cloudy weather; one can also see how the PPFD changes from sunny to cloudy days at different times of the year.

China and United States, temperature and sunshine

10 December 2015

My column in the November/December 2015 issue of GCM China shows normal temperature and sunshine data from three pairs of cities. I chose Beijing and Philadelphia to represent the cool-season zone, Shanghai and Atlanta to represent the transition zone, and Guangzhou and Tampa to represent the warm-season zone.

In each of these pairs, the temperatures are similar. In the transition zone and warm-season zone pairs, the sunshine hours are quite different, and it is the locations in the USA with higher sunshine. From the sunshine hours, one can estimate the daily light integral (DLI). As I wrote in the article:

Each sunshine hour is equivalent to about 5 moles of photons per square meter. Taking the month of May as an example, on a normal day in Guangzhou, there would be 4 hours of sunshine, in Tampa 10, in Shanghai 5.5, and in Atlanta 9. This gives an approximate DLI of 20 in Guangzhou, 50 in Tampa, 27.5 in Shanghai, and 45 in Atlanta. Bermudagrass grows best when the DLI is above 40.

Read the full article here in Chinese or in English.

A chart of PPFD at two locations this year from January 1 through last Friday

25 October 2015

The photosynthetically active radiation (PAR) changes through the day and through the year. The PAR is measured instantaneously for a duration of 1 second as the photosynthetic photon flux density (PPFD), and by adding up the PPFD for all the seconds in the day, one gets the daily total of PAR, which is called the daily light integral (DLI).

These charts show the average PPFD on an hour by hour basis. With a look at a chart like this, one can see:

• how length of the day affects PAR, by looking at what time in the morning and what time in the evening the PPFD goes to 0.
• how time of the day affects PAR, by looking at the change in PPFD hour by hour through the day.
• how day of the year, and consequently sun angle, affects the PAR, by looking at the maximum values of PPFD at midday and seeing how they change through the year.
• how clouds reduce the PAR, by comparing PPFD on sunny hours or days to PPFD on hours or days that don't have full sun. For more about sun and clouds and time of year, see these descriptive slides with data from 4 days in Tokyo this year: a sunny summer day, a very cloudy summer day, a sunny autumn day, and a partly sunny autumn day.

This chart shows, for every hour of this year through last Friday, the average PPFD for that hour at Tokyo (red) and at Watkinsville (blue). Each panel of the chart is a single day, and the DLI in units of mol m-2 d-1 is written on each panel, in red for Tokyo and in blue for Watkinsville.

There have been 296 days this year, through October 23. On one of these days, February 10, there were erroneous data at Watkinsville, so I don't have a DLI. That leaves 295 days with a DLI for both Tokyo and Watkinsville. These locations have similar temperatures, and similar latitudes. How do they compare for photosynthetically active radiation? There have been 115 days with a higher DLI at Tokyo than at Watkinsville, and 180 days with a higher DLI at Watkinsville than at Tokyo.

I've made a couple other similar charts. This one shows the average PPFD at Tokyo hour by hour this year through October 12. Because the chart shows data for only one location, I've used color to indicate the month.

And the next one is the same location and dates as the above, with the addition of the DLI written on each panel.

The Watkinsville data are from the US Climate Reference Network and the Tokyo data are from the Japan Meteorological Agency.

Bentgrass in hot and not so hot places

19 October 2015

Creeping bentgrass is a cool-season grass. When temperatures are hot, it doesn't perform well. I was asked if bentgrass in southern China was comparable to bentgrass in Spain. I don't think that is the right comparison. It would be more appropriate to compare southern China to Florida.

I downloaded the 2014 daily temperatures for the international airports at the cities shown in this chart, then plotted the cumulative sum of the mean temperature for the year.

Guangzhou and Orlando had the same cumulative sum of temperature. Bentgrass wouldn't be a good choice in Orlando, and I don't think it is a good choice in Guangzhou either.

A better way to look at bentgrass suitability is to look at the low temperatures. If the low temperatures are too high, for too many days, bentgrass will be really difficult to manage, eventually becoming too much of a problem and one would be better off with a warm-season grass.

For 2014, here's the number of days with a low temperature greater than or equal to 22°C. I'd look at anything more than 60 days in a year above that level as being difficult for bent.

This way of evaluating the temperature fits pretty well how one expects bentgrass to perform in these locations. Perfect in Kunming, the "Spring City." Pretty good in Spain. A challenge in Shanghai summers, with some warm-season greens there also, but possible with good management. Not used in Orlando. And I wouldn't want to try it in Guangzhou.

For more about temperatures and bentgrass, see:

The data prove me wrong

06 October 2015

I was listening to Frank Rossi talking with Bert McCarty on the TurfNet Radio Network. It was an interesting conversation, about lots of things including ultradwarf bermuda, pigments, and heat stress on bentgrass.

One thing struck me when they were talking about ultradwarf bermudagrass, and Rossi mentioned that "light levels change, maybe temperatures change, but certainly light levels is a driving force" in the slowdown in growth at the end of summer. McCarty confirmed that "light starts it off" as the "grass starts to slow up" even though it is still hot.

I wondered if that was right. It seemed to me that temperatures would go down to slow growth before light would. Well, I looked up some data, and they were absolutely right.

Here's the average monthly temperature for the last 10 years at McClellanville, South Carolina. Temperatures for 2015 are included to October 4. Looks like July and August are the hottest months, no surprise, and I see a drop going to September.

Then I looked up the global solar radiation, converted to photosynthetically active radiation expressed as a daily light integral, and calculated the monthly averages. Again, 2015 data are just up to October 4. The DLI peaks in June and then looks like a steady decrease to December.

Ok, I had the temperature and the light, but how to see which one is dropping off at a different time? To do that, I looked at the change from month to month. I calculated something called the log percent (L%), which is $100(log_{e}(y/x))$. In this case I let $y$ be the value for the month, and $x$ be the value for the previous month. This calculation gives a symmetric, additive, and normed measure of the relative change in light, and of the relative change in temperature.

I didn't compare December to January, so there is no L% for January. That's alright, because I was most interested in what happened at the end of summer, specifically in August and September. The values on this chart for February represent the L% change from January, March shows the change from February, and so on.

It is really clear. Dr. McCarty was exactly right.

In July the temperature is still increasing from the previous month, and light is decreasing from its peak in June.

Then in August there is a big drop in light, and on average no change in temperature. And even in September, the relative drop in light is higher than is the relative drop in temperature. It is not until October and November that the temperature decrease is more than the light decrease.

It's good to know this!

Energy for growth, and weeds

29 September 2015

Two things today are kind of related to this topic. One is this -- Jim Brosnan mentioned, and showed photographic evidence, that "weed pressure on Oahu never ceases to amaze."

And I had a conversation with a golf course designer about fine fescue as an infrequently mown rough, in what climates that species can work, and what happens when it is too hot for fine fescue. And I mentioned that one can plant a number of species other than fine fescue in a warmer climate, but the problem becomes one of "how can we find a ball" because there is a lot of energy for growth. Of course there are various techniques turf managers can use to solve that problem, but then the turf will be alive, but thin. It must be if one is going to find a ball in it.

Once there are voids, weeds have an opportunity to grow. Turf managers can solve this problem too, with herbicides, or with manual removal of weeds. But now comes another problem. That is erosion, in locations with substantial rainfall.

Anyway, it must be that the growth of plants (desired species, and weeds) is related to the energy available for the plants to grow. In general the hotter it is, the more energy there will be for weeds, so when one thins a low maintenance rough, the energy for weed growth or invasion is going to be more in a hotter climate than in a cool one. I looked up some data from Japan -- hour by hour data of temperature and global irradiance for 2014 at Sapporo, Tokyo, and Naha. Then I converted the irradiance to photosynthetically active radiation (PAR) using a factor of 2.04.

I looked only at day time, when the sun was above the horizon. And I arbitrarily cut the data to look only at those hours when the temperature was greater than or equal to 20°C. Then I added up all the light, and all the hours. This is a very rough index of how much energy there is for growth, especially for the weeds that will grow when it is hot. And it is a conservative estimate, because the night temperatures influence growth too, and so does the actual temperature. This is just a quick way to note the differences between locations.

At Sapporo in 2014, the cumulative sum of PAR for hours when the temperature was greater than or equal to 20°C was 3,781 mol m-2. Tokyo has 5,844, and Naha was 9,124. Oahu is considerably warmer than Naha, so it almost certainly would have more PAR than Naha at this cutoff value.

Just looking at the time, how many hours were there for weeds to grow well in these different locations, by looking at how many hours there were with a temperature at or above 20°C? At Sapporo, there were 1,365 of these hours; at Tokyo there were 2,503; and at Naha it was 3,805. Again, locations in Oahu would almost certainly be more than Naha.

That is a real quick estimate of how much energy there is for weeds to grow, or more specifically how the energy is likely to differ in magnitude from location to location.

And one more thing -- in Scotland where a fine fescue rough actually works well, how much energy would there be for weeds? I don't have the exact irradiance data for Scotland, so I won't try to compare it to exact measurements. But I can give some idea of just how much lower the energy is, or how much lower the duration of time would be for weeds to grow rapidly. Huge disclaimer is necessary here, because the species are different, so a C3 weed like Poa annua might grow relatively rapidly in Dornoch but I am considering more the C4 weeds like Paspalum dilatatum or Cyperus rotundus.

It still makes an interesting comparison. Of Naha, Tokyo, and Sapporo, Sapporo is by far the coldest. And in the hottest month of the year in Sapporo, the average low temperature is 19.1°C, and the average high temperature is 26.4°C. How about somewhere in Scotland where fine fescue grows well? I picked Leuchars, just north of St. Andrews. In the hottest month of the year in Leuchars, the average low is 10.8°C, and the average high is 19.2°C.